Skip to main content

Full text of "An experimental enquiry concerning the natural powers of water and wind to turn mills"

See other formats


This is a digital copy of a book that was preserved for general ions on library shelves before il was carefully scanned by Google as part of a project 

to make the world's books discoverable online. 

Il has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject 

to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books 

are our gateways to the past, representing a wealth of history, culture and knowledge that's often diflicult to discover. 

Marks, notations and other marginalia present in the original volume will appear in this file - a reminder of this book's long journey from the 

publisher to a library and finally to you. 

Usage guidelines 

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the 
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing this resource, we have taken steps to 
prevent abuse by commercial parlies, including placing technical restrictions on automated querying. 
We also ask that you: 

+ Make non-commercial use of the plus We designed Google Book Search for use by individuals, and we request that you use these files for 
personal, non-commercial purposes. 

+ Refrain from automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine 
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the 
use of public domain materials for these purposes and may be able to help. 

+ Maintain attribution The Google "watermark" you see on each file is essential for informing people about this project and helping them find 
additional materials through Google Book Search. Please do not remove it. 

+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just 
because we believe a b<x>k is in the public domain for users in the United States, that the work is also in the public domain for users in other 

countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of 
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means il can be used in any manner 
anywhere in the world. Copyright infringement liability can be quite severe. 

About Google Book Search 

Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers 
discover the world's hooks while helping authors ami publishers reach new audiences. You can search through I lie lull text of this book on I lie web 
at |http : //books . qooqle . com/| 



^ te-ivubd { A fate.. ti^.^ct{- S-*/ ?A 

itj. -uZa-£ •*>.„* 




Experimental Enquiry 



O F 



Turn MILLS, and other Machines, 
depending on a circular Motion. 



By j/SMEATON, F.R.S. 1 4-n^v 

\ ii ' 


Printed in the Year M.dcc.lx. 

v *- 


A N 

Experimental ENQUIRY 



O F 

W A T E R and WIND 

T O 

Turn MILLS, and other Machines, 

depending on a circular Motion. 

*«'"«n. TITHAT I have to communicate on 
*■«, .759- yy this fubject was originally deduced 
frora experiments made on working models, which 
I look upon as the beft means of obtaining the out- 
lines in mechanical enquiries. . But in this cafe it 
is very neceuary to diJiinguHh the circumftances in 
which a model diners from a machine in large; 
otherwife a model is more apt to lead us from the 
truth than towards it. Hence the common obferva- 
A a tion, 


• " " E 4 •} • ^ ■ ■ 

tlon, that a thing may do very well in a model,, that 
will not anfwer in large. And indeed, tho' the ut- 
moft circumfpeftion be ufed in this way, the beft 
ftrudture of machines cannot be fully afcertained,. 
but by making trials With them, when made of their 
proper fize. It is for this reafon, that, tho' the mo- 
dels referred to, and the greateft part of the follow- 
ing experiments, were made in the years 1772 andr 
1 7 f 3 , yet I deferred offering them: to the Society; 
till I had an opportunity of putting the deductions 
made therefrom in real practice, in a variety of cafes* 
and for various purpofes ; fo as to be able to aflure: 
the Society, that I have found them to anfwer. . 

PARt I. 
Concerning Undershot Water-Wheels. 

Plate IV. Fig. 1. is a perfpedive view of the ma- 
chine for experiments on water-wheels ; wherein 

A BCD is the lower cittern, or magazine, for re- 
ceiving the water, after it has quitted the wheel ;, 
. and for fupplying 

D E the upper cifteni, or head 5 wherein the water 
being railed to any height required, by a pump, 
that height is fliewn by 

F G, a fmall rod, divided into inches and parts j. 
with a float at the bottom, to move the rod up 
and down, as the furface of the water riles and' 
falls. . 

H I is a rod by which the fluice is drawn, and ftopt 
at any height required, by means of 
K a pin or peg, which fits feveral holes, placed 


[ s ] 

In the manner of a diagonal (bale, upon the face 
of the rod HI.. 

G L is the upper part of the rod of the pump, for 
drawing the water out of the lower ciftern, in 
order to raife and keep up the furface thereof 
at its defired height, in the head DEj thereby 
to fupply the water, expended by the aperture 
of the fluke. 

M M is the arch and handle for working the pump, 
which is limited in its ftroke by 
N a piece for flopping the handle from railing the 
pifton too high $ that alfo being prevented from 
going too low, by meeting the bottom bf the 
O is the cylinder, upon which a cord winds, and 
which being conducted over the pulliea P and 
Q^ raifes 

- R, the fcale, into which the weights are put, for 
trying the power of the water. 

,S T the two ftandards, which fupport the wheel, 
are made to Aide up and down, in order to ad- 
juft the wheel, as near as poffible, to the floor 
of the conduit; 
W the beam which fupports the fcale and pul- 
lies ; this is reprefented as but little higher than 
the machine, for the fake of bringing the figure 
into a moderate compafs, but in reality is placed 
I f or i & feet higher than the wheel. 

Plate V. Fig. a. is a fe&ion of the fame machine, 
wherein the fame parts are marked with the fame 
letters as in Fig. i. Befides which 
X X is the .pump barrel, being 5 inches diameter, 
and 1 1 inches long* 

Y is 

V _ 


Y is the pifton ; and 

Z the fixed valve. 
GVija cylinder of wood, fixed upon the pump- 
rod, and reaches above the furface of the wa- 
ter ; this piece of wood being of fuch a thick- 
nefs, that its feftion is half the area of that of the 
pump-barrel, will caufe the furface of water to 
rife in the head, as much while the pifton is 
descending, as while it is rifing : and will there* 
by keep the gauge-rod F G more equally to its 
height. Note, the arch and handle M M is here 
reprefented on a different fide to what it is ihewn 
in the preceding figures, in order that its dim en - 
lions may the better appear. 

a a £bews one of the two wires which ferve as di- 
rectors to the float, in order that the gauge rod 
F G may be kept perpendicular ; for the fame 
purpofe alfo ferves w, a piece of wood with a 
hole to receive the gauge-rod, and keep it up- 
. h is the aperture of the fluice. 

c c a kant-board, for throwing the water more di- 
reftly down the opening c d, into the lower 
ciftern: and 

c e is a Hoping board, for bringing back the water 
that is thrown up by the floats of the wheel. 

Fig. $, represents one end of the main axis, with a 

fe&ion of the moveable cylinder, marked O in the 

preceding figures. 

ABCD is the end of the axis j whereof the parts 

B and D are covered with ferrules or hoops or brafs. 

£ is a cylinder of metal ; whereof the part marked 



F is the pivot or gudgeon. 

cc is the feftion of an hollow cylinder of wood, 
the diameter of the interior part being fome- 
what larger than the cylindrical ferrule B. 

a a is the fe&ion of a ferrule of brafs, driven into 
the end of the hollow cylinder, and which is 
adjufted to that marked B, fo as to Hide freely 
thereupon, but with as little (hake as poffible. 
bb, dd,gg, reprefent the fe&ion of a brafs ferrule, 
plate, and focket, fixed upon the other end of 
the hollow cylinder} the focket dd being ad- 
jufted to Aide freely upon the cylinder E, in the 
iaroe manner as the ferrule *a Hides upon the 
. cylinder B : the outer end of the focket at 

g g is formed into a fort of button ; by pufhing 
wfeeneof^ the hollow cylinder will move back- 
wards and f oiwacds, or tarn round at pleafcre 
upon the cylindrical parts of the axis B and £• 
e * y i r, », icprefent the fefition of a brafe ferrule, al- 
fo fixed upon the hollow cylinder: the edge of 

-*e: i$ cut into teeth, in. the manner of k tcntratt 

wheel f and the edge thereof 
v * is cut in the manner of a ratchett* 

' Of ccaifcqocnce, when the plate bddi h 
puflied clofe to the ferrule D, the teeth of the 
ferrule * e will lay ihctfd of 
G, a pin fixed into the axis ; by which means the 
hoflow cylinder as made to turn along with the 
wheel and axis:: but being drawn back by the 
button gg y the hollow cylinder is thereby, dif- 
♦engaged .from the pin G, and ©eafes turning. 
JV&fr» The weight in the icale is prevented 



from running back, by a catch that plays in 
and lays hold of the ratchet < o<>. 

By this means the hollow cylinder upon which 
the cord winds, and raifes the weight, is put in 
ad: ion and difcharged therefrom inftantaneoufly, 
while the wheel is in motion : for without fome 
contrivance of this kind, it would not be eafy 
to make this fort of experiments with any to- 
lerable degree of exa&nefs; 

• The ufe of the . apparatus now defcribed will be 
rendered more intelligible, by giving a general idea 
of what I had in view ; but as I fhall be obliged to 
make ufe of a term which has heretofore been the 
caufe of difputation, I think it neceffary to affign the 
fenfe in which I would be uuderftood to ufe it ; and 
in which I apprehend it is ufed by practical Mecha- 
nic ks. 

The word Power? as ufed in pradical mechanicks, 
I apprehend to fignify the exertion of ftrength, , gra- 
vitation, impulfe, or prefTure, fo as to produce mo- 
tion: and by means of ftrength, gravitation, im- 
pulfe, or preflure, compounded with motion, to be 
capable of producing an effedt : and that no effect is 
properly mechanical, but what requires fuch a kind 
of power to produce it. 

The railing of a weight, relative to the height to 
which it can be raifed in a given time, is the moil 
proper meafure of power ; or, in other words, if the 
weight raifed is multiplied by the height to which 
it can be raifed in a given time, the produdt is the 
meafure of the power railing it ; and confequently, 
all thofe powers are equal, whofe products, made by 



[ 9 ] 

foch multiplication, are equal: for if a power can 
raiie twice the weight to the fame height ; or the 
Tame weight to twice the height, in the fame time 
that another power can, the firft power is double the 
fecond : and if a power can raife half the weight to 
double the height j or double the weight to half the 
height, in the fame time that another can, thofe two 
powers are equal. But note, all this is to be under- 
ftood in cafe of flow or equable motion of the body 
raifed ; for in quick, accelerated, or retarded mo- 
tions, the vis inertia of the matter moved will make 
a variation. 

In comparing . the effects produced by water- 
wheels, with the powers producing them ; or, in 
other words, to know, what part of the original 
power is neceffarily loft in the application, we mull 
prev&oufly know now much of the power is fpent 
in overcoming the fridion of the machinery, and 
the refiftance of the air $ alfo what is the real velo- 
city of the water at the inftant that it ftrikes the 
wheel ; and the real quantity of water expended in 
a given time. 

, From the velocity of the water, at the inftant that 
it ftrikes the wheel, given ; the height of head pro- 
ductive of fuch velocity can be deduced, from ac- 
knowleged and experimented principles of hydrofta- 
tics : fo that by multiplying the quantity, or weight 
of water, really expended in a given time, by the 
height of head fo obtained ; which muft be confi- 
dered as the height from which that weight of wa- 
ter had defcended in that given time j we ftiall fiave 
z product, equal to the original power of the water y 
"and clear of all uncertainty, that would arife from* 
the fri&iori of the water, in paffing fmall apertures ; 

B ; * r and 


and from all doubts, arifing from the different 
fure of fpouting waters, afligned by different authors. 
On the other hand, the fum of the weights raifed 
by the a&ion of this water, and of the weight re- 
quired to overcome the fri&ion and refiftance of the 
machine, multiplied by the height to which the 
weight can be raifed in the time given, the product 
will be equal to the cfFed of that power ; and the 
proportion of the two products will be the proper-* 
tion of the power to the effefi : fo that by loading 
the wheel with different weights fuccefHycly, we 
ihall be able to determine at what particular load* 
and velocity of the wheel* the effeft is a maximum. 

The manner of finding the real velocity of the 
water, at the inftant of its ftriking the wheel ; the 
manner of finding the value of the fri&ion, refin- 
ance, &c. in any given cafe; and the manner of 
fading the real ex pence of water, fo far as cojok 
eerns the following experiments, without having re- 
course to theory ; being matters upon which, the fol- 
lowing determinations depend, it will be neceflary 
to explain them* 

To determine the Velocity of the Water Jhrihkng the 


It has already been mentiened^ in the reference* 
ta the figures* that weighta are raifed by a cord 
winding round a cylindrical part of the ax$. Firft, 
then, let the wheel be put in motion by the water, 
but without any weights in the fcale h and let the 
number of turns in a minute be 60 1 now it is evi ? 
<knty that was the wheel free from fridion and refin- 
ance,, that 60. times the circumference of the wheel 

:• •: ....:: Would 

• » 

[» 3. 

would be the fpace through which the water would 
have moved in a minute * with that velocity where- 
with it ftruck the wheel: but the wheel being in- 
cumbred by fri&toti and refinance, and yet moving 
60 turns in a minute, it is plain, that the velocity 
of the water muft have been greater than 60 cir- 
cumferences before it met with the wheel. Let 
now the cord be wound round the cylinder, but 
contrary to the ufual way, and put a weight in 
the fcale ; the weight fo difpofed (which may be 
called the counter-weight) will endeavour to affift the 
wheel in turning the fame away, as it would have 
been turned by the water : put therefore as much 
weight into the fcale as, without any water, will 
caufe it to turn fomewhat fafter than at the rate of 
60 turns in a minute ; fuppofe 63 : let it now be 
tried again by the water, affifted by the weight ; the 
wheel therefore will now make more than 60 turns; 
fuppofe 64 : hence we conclude the water ftill ex- 
erts feme power in giving motion to the wheel. Let 
the wfeight be again increafed, fo as to make 6\\ 
turns in a minute without water : let it once more 
be tried with water as before ; and fuppofe it now 
to make the fame number of turns with water as 
without, viz. 64I : hence it is evident, that in this 
cafe the wheel makes the fame number of turns in 
a minute, as it would do if the wheel had no fric- 
tion or refiftance at all j becaufe the weight is equi- 
valent thereto 5 for was it too little, the water would 
accelerate the wheel beyond the weight $ and if too 
great, retard it ; fo that the water now becomes a 
regulator of the wheel's motion ; and the velocity 
of its circumference becomes a meafure of the vclo* 
city of the water, 

B a In 

C " 3 

In like manner, in feeking the greateft produfl; 
or maximum of effedt \ having found by trials what 
weight gives the greateft produdt, by limply multi- 
plying the weight in the fcale by the number of turns 
of the wheel, find what weight in the fcale, when 
the cord is on the contrary fide of the cylinder, will 
caufe the wheel to make the fame number of turn9 
the fame way, without water $ it is evident that this 
weight will be nearly equal to all friction and refin- 
ance taken together; and confequently, that the 
weight in the fcale, with twice * the weight of the 
fcale, added to the back or counter- weight, will be 
equal to the weight that could have been raifed, fup- 
pofing the machine had been without friction or re- 
finance 5 and which multiplied by the height to 
which it was raifed, the product will be the greateft 
efFedt of that power. 

. The quantity of water expended is found thusr 

The pump made ufe of for replenishing the head, 
with water was fo carefully made, that no water 
efcaping back by the leathers,., it delivered the fame 
quantity of water at every ftroke, whether worked 
quick or flow ; and as the length of the ftroke was 
limited, confequently the value of one ftroke (or on 
account of more exadtnefs I a ftrokes) was known, by 
the height to which the water was thereby raifed in 
the head ; which being of a regular figure was eafily 
meafured. The fluice, by which the water was drawn 
upon the wheel, was made to flop at certain heights 
by a peg ; fo that when the peg was in the fame hole, 

* The weight of the fcale makes put of the weight both ways* 

8 the 


the aperture . for . the effluent water was the fame. 
Hence the quantity of wates expended by any given 
head, and opening of the fluice, may be obtained : for 
by obferving how many ftrokes a minute was fuffi- 
cient to keep up the furface of the water at the given 
height, and multiplying the number of ftrokes by the 
value of each, the water expended by any given 
aperture and head in a given time will be given. 

Thefe things will be further illuftrated by going 
over the calculus of one fett of experiments. 

Specimen of a Sett of Experiments. 

The fluice drawn to the i ft hole. 
The water above the floor of the fluice 30 Inches. 
Strokes of the pump in a minute — 397 
The head raifed by 12 ftrokes — 21 Inches. 
The wheel raifed the empty fcale, and made Qirns 

in a minute ■ ■ 80 

With a counter- weight of 1 lb. 8 oz. it made 85 
D° tri^d with water — — \ — 86 

N° Weight. Turns' in a mm. Produft. . 

x — , 4 o — 45 — 180 
2 — ' 5 6 — > 42 — 210 
3, — 60 — 36% — 217! 

4 — 7 o — 33* — 2361 

5 — 8 o — 30— 240 maximum^ 

6 — 9 o — 261 — 238$ 

7 — 10 o — 22 — 220 

8 — 11 o — 16} — 181* 

9 — 12 * ceafed working. 

* N* B. When the wheel moves fo flbw as not to rid the wa- 
ter fo faft as fupplied by the fluice, the accumulated water falls 
back upon the aperture, and the wheal immediately ceafes moving. 



Counter-weight, for 30 turns without water, a oz. in 

the fcale. 
•J/. B. The area of the head was 1 05,8 fqoare inches. 

Weight of the empty fcale and pulley, 10 oz. 

Circumference of the cylinder, 9 inches. 

Circumference of the water-wheel, 7$ ditto. 


Reduction *f the above Sett tf Experiments. 

The circumference of the wheel, jf inches, mul- 
tiplied by 86 turns, gives 6450 inches for the velo- 
city of the water in a minute 5 ^~ of which will Be 
the velocity in a fecond, equal to 107,5 Inches, or 
8,96 feet, which is due to a head of 1 f inches * ; 
and this we call the virtual or effe&ive head. 

The area of the head being 105,$ inches, this 
multiplied by the weight of Water of the inch cubic, 
equal to the decimal ,579 of the ounce avoirdupoife, 
gives 61,26 ounces for the weight of as much water, 
as is contained in the head, upon 1 inch in depth, 
tV of which is 3,83 pounds ; mis multiplied by the 
depth 2 1 inches, gives 80,43 lb. for the value of 1 2 
ftrokes ; and by proportion, 3p| (the number made 
in a minute) will give 264^7 M?. the weight of wa- 
ter expended in a minute. _ 

Now as 264,7 lb* of water p&ay be coniidered as 
faviftg defended through a ipace of 1 5 inches in a 
minute, the prodoft of thefe two numbers 3970 will 
"exprefs the **w*r of the water to produce mechanical 
effects ; which were as fQllows. 


. * This, is det e rm ined upon the co mmon ma x im of h y d r o fl aticg» 
that the velocity of flouting; waters is equal to th* vilocky that 
an heavy body would acquire in falling from, the height of the 
refejVQir 1 and i* proved by the lifog ofjets to. the height of their 
vefervoift nearly. 


The velocity of the wheel at the maximum, as ap- 
peals above, w&* 30 turns a minute ; which mul- 
tiplied bv$> inches, the circumference of the cylin- 
der, makes 270 inches $ but as the fcale was hung 
by a pulley and double line, the weight was only 
raifed half of this, viz. 135 inches. 

The weight in the fcale at the maximum $16. ooz. 

Weight of the fcale and pulley — — p 10 

Counterweight, fcale, and pulley o 12 


. Sum of the refiftance 9 6 

ox lb. 9>57S* 
Now as 5>>375 ltx is raifed 13 J inches, thefe two 
numbers being multiplied together, the produd is 
1266, which exprefles the eflfedt produced at a ma- 
ximum : fo that the proportion of the power to the 
effeB is as 3,970 : 1266, or as 10 : 3,18. 

But tho this is the greateft^g/i effect producible 
from the power mentioned, by the impulfe of the 
water trpts* aa andcrihot wheel ; yet, as the whole 
pernor of the water is not exhausted thereby, this 
will not be the true ratio between the fwkr of the 
water, and the ftm of all the effe&s producible 
thendfcooi. : for as the water muft acceflarily leave the 
wheel wkto a velocity equal to the wheel'* circum- 
ference, it is plain that fome part of the power of 
the water txxaSt remain after quitting, the wheel. 

The velocity of the wheel at the maximum is 30 
turns a mhwte ; and confeqocntly its circumference 
moves at the rate of 7^1 2 5 feet a fecond, which an- 
fwers to a, head i,£i inches $ this bring multiplied 
by the expence of water in a minute, ofz. 2*64,7 lb; 
produces 481 for the power remaining in the water 
after it has pafled the wheel: this bemg therefore 


[ i«] 

deduced from the original power 3970, leaves 3489, 
whidh is that part of the power which is fpent in 
producing the efFedt 1266; and confequently the 
part of the power fpent in producing the effect, is to 
the greateft efFe<3: producible thereby as 3489 : 1266 
:: 10 : 3,62,or as 11 to 4. 

• The velocity of the water ftriking the wheel has 
been determined to be equal to 86 circumferences of 
the wheel per minute, and the velocity of the wheel 
at the maximum to be 30 $ the velocity of the water 
will therefore be to that of the wheel as 86 to 30, or 
as 10 to 3,5, or as 20 to 7. 

The load at the maximum has been fhown to be 
equal to 9 lb. 6 oz. and that the wheel ceafed move- 
ingwith 12 lb. in the fcale : to which if the weight 
of the fcale is added, viz. 10 ounces *, the propor- 
tion will be nearly as 3 to 4 between the load at the 
maximumznd that by which the" Wheel is flopped. 

It is fomewhat remarkable, that tho' the velocity 
of the wheel in relation to the water turns out greater 
than -j of the velocity of the the water, yet jthe im- 
pulfe of the water in the cafe of. a maximum is more 
than double of what is affigned by theory $ that is, 
inftead of | of the column, it is nearly ^qual to the 
whale column. 

It muft be remembred, therefore, that, in the pre- 
fent cafe, the wheel was not placed in an open river, 
where the natural current, after it has communicated 
its impulfe to the float, has room on all fides to es- 
cape, as the theory fuppofes ; but in a conduit or 

f * The refiftance of the air in this cafe ceafes, and the fri&ion 
is not added, as 12 lb. in the fcale was fufficient to flop the wheel 
after it had been in full motion ; and therefore fomewhat more 
than a counterbalance to the impulfe of the water, 



E 17] 

race, to which thq fi^ftt }>c«p§ adapted, the water 
cannot other wife efcape than by moving along with 
{he wheel. It is obkfrvablq, that a wfieel working 
in this munnff, as foop as the water meets the float* 
receiving a fudd$n chepk, it fifes up agpinft (the flpat* 
like a w*ve ag^inft 3 fixed objeft $ infomuch that 
when the ihest <>f wader is not a quarter pf an ipcb 
thick before it pietts the fkgt, vet this Iheet wjll 
aft upon the whple furfaee of a float, iwhefe height 
is g inchqs ; said jconfequently was tibe flo^t qo higher 
than tto tfctckncjfs of die ihqet of waft?, as the 
cry alfo foppofes; a gteat j>art of the force wquid bare 
feet* lpft> hy tfae water's jdafhing over the £oat *. 

to fortfier confirmation of what is already deli- 
vofcd, J have joined the following tabl^ ,cQntaiiv* 
»g the *efu*k <n 27 fetts of (experiments, j^ade ^nd 
reduced in She ,pwtnj?e|r above ipeqiffcfl. ^h# jre*- 

rally 'fbflojW ftom a cqmpswjfon c£ the different ex- 
patimenfts tQgc&er, 

* Since ^he al)oye >vas ,v?r0te, J : find that Prpf<iflQr Eiiler^ip tfc 
Bedin A&s for thenar 1348^ in a jpemoire intitfied, fitaxims tour 

> which >fr cm to jbe; tie more reraariaUc, >as I qoat nnd ,bejha* 

>gbp*n4*y4)^0frtiation of the principle tfctrein contained* either 

&t^ ^ovy cfr-eKpetimertt; *>rthas mode ,atiy ufe thereof in h» 

'ctffcmtottons cm this fiibj^b.— ^ dependant dans ce <faa puifque 

"** l^au eft r^ltechic, |& quelle fccoule Tur les allies vers Tes cqcEs, 

■*" efie y exsree MCpre une force pa*tictt%e, flont r«$t de l^ni- 

; " pulfionfeia aqgtxujite; &; experience jointer la tfcteorie a fait 

" 4C *oir jjue 4%»s .ce cas # Ja force sjfc .prefijue. double ibue 

u qa'il Japt prendre le.4wi^le <te le fc&$a du til d'eaupour -oe 

M :i|ui r rfKp«nd ! dae3CQcas^a je furfeoe des aubes, pourvu.igu'fUe» 

*~ fci c nt a ffte largto pour rcc evoir ce fupplem e nt ^e forqe. Cfrfi 

^ Its aube»t|ite>iem -plus targes que lefi], on t*ak d'eau on ne 

m4A ik vro it pieiidie qu e nc finale f e dj on, tQutcorrrrne dans le pre-^ 

M Aier^s, -on-Paube totfte entire eft pappee par Tesn^ 



[ 18] 













10 : 3.5 

!° : 3»4 


,Q :3^7 







10; 7,32 

10:8,02 lift 

10:8,3 /bole, 





* ■ 















1 o : 2,99 


i° : 3.i3 







' f P ; 4>5 5 



















r 19] 


Maxims and Obfervations deduced from the foregoing 

Table of Experiments. * 

Maxim I. That the virtual or effeftinje bead being 
the fame, the effeSt will be nearly as the quantity of 
water expended. 

This will appear by comparing the contents of the 
columns 4, 8, and io > in the foregoing fetts of ex- 
periments $ as for 

Example iji f taken from N°. 8. and 2$, viz; 

No. Virtual Head. Water expended.. • E0e& 
8 7,29 — rr- l6i - — — r* 328 / 

a-f 7,29 355 — 785 

Now the heads being equal $ if the eflecffs are pro- 
portioned to the water expended, we fhall have by 
maxim ift, 161 \ 357 :: 328 : 723* but 723 falls 
lhort of 78 57 as it turns out in experiment according 
to N°. 2 j, by 62 ; the effeft thertfofre of *J°. 25, 
compared with N°. 8, is greater tharr according to 
the prefent maxim in the ratio of 14 KrT3. ' 
: The foregoing example, wj^h four firnilarjqnes, 
are feen at one view m the following Tables : - , .. , f 






W o 


I .. 




2 '5 








O f— « e 

CU <* Q 

O 5 *■» 

h O « 

357:: 975:1221 

«624» 4 

H: 13 




1 0,5 


I2IO| * 


3d I 



2 SS 

J86 fc** ; 3 * 2 :: 541 :7 ° 4 


3* : 391 





|g^|223:262:4 3i7t564 


48-:.'! 7 




S34 J 3 ° 7 : 36 ° : " 45 ° : 531 3 +| I78:l 77 

C 2 


Hence therefore, in comparing different cxperi- 
Ifnerits, las fome fall ihorL and others exceed die 
maximum, and all agree therewith, as jiear as can 
fee expfe&ed, in tti affair where lb many different 
tii^flmttsftices are concerned 5 we may, according to 
the laws of reafoning by indu&ion, conclude the 
tnaifrft ttut 3 01 z. that ttie effefts are nearly as the 
'^uafitity df ^dttw: feXpflfded. 

M&kh A. 7&tf -ft&fr taprtir** <ef ' <wafer teing the 
fate, Hbe eJpS? «*»// 4i nearly *s *ke height *f the 
vif>m l vr effeai ik>beail. 

* » • • 

'Thiis alio will appear by comparing the contents 
of 'columns 4, \ 2nd it>, la any of the fetts of ex- 

: Jixaptfk •#> ^-W. a, and N°. 24. ««?. 

. • ,N\ V5tt.tkfd. E^«ce. £&&. 

-ft* — • *5 ■' — 864,7 " *a$6 

24 _ 4,7 ^— . *6 a ,. „•,„„ , 3 .8f 

"Notir as&e^expewdis art taw quite equal, wemuft 
proportion one <* l!he fcfifccHs accordingly r 'thus 
Iby maxim ~Yft;~ a5a : 5264,7 :: '.jU j : £89 
land by max. ad, jj : 4,7 :: n6o" : 397 

Difference ^m , ■ 18 
ie ie&&lthqrefere of~N°. 4:4. oompaml with N°. 
•».4s lefs t han aec or din g-to the j ri fi^ 4U4xim ?a-tfap 
tiitoeff '4^:50. 

'The fptegoing, and two -other JtmLiaregafOplcf, 
are comprised in the fallowing Table. 

I ** p 


-Comiarifon. • 





*H< «#* 



Sr i*» 

OIL ..I 

1266 1 Mpx. ift, 262 ^64,7 :: 385 ^ 319 7 a I , 
385J i Mjix. 2d, a 5 : .4,7 :: 1266 5 3^1* f "lf v ' * 

}M«.4i y Jil-;^5 ;:ai7 4 ^1 ^. 
jl^x.2fti 5 & 53,5^: 14^:316 J *+ 

■ 4 i » * — r— - r-s— f = — ; — j - \ »i — ■ ■ ■ — ; 
*iA 3f* i5<>5 IMf*- •* *^S;« 34* :: *42 : 433 1 . fi 

{MStf 6 h P} 2 fMfi*' & 'tf^'P? '5J5 : 45°f I 7 V s V 

8 : 9 

1 . ' * 

Ma^injL III. TJiaf the quality jof wafer txpendej 
km m g the fim* n the <$*& it ty(uJ)L,as tpe /qua re of 

.. This jviJJ appear by CQpiparinff the confcnts of 
,^9lV!!mns : 3,^^iii fro, in : ^y,of *he Xetfs ©f ^xpcri. 
jgeots; as jfbr 

fiy^r |£ gf N*. * **/** **. 2,4. «fe. 

. ^T. Tuwis m a,mi«. -Expeiy*. Efieft. 

a < i w ftf .— * — p 2^4,7 -w-j — ia66 
>4 .*- 1 — ^ $* .' i" y 2j62 -^--t — 38f 

^e y^lcjcity^lng ;as Ae nmplber ojT turns, welhaljl 

type, ■■;■ " ^ " / 

. ;bynnx.4:ft, *$2 ,; 204,7 :: 38^ : 3*9 

Tliflf*»ri>n/*» ••.•."' r 

TEhej^nVt gterejpse of N c . 24. compared with N°. -than-by .the prefent maxim in _the.iatio jof 

3?he foregoing, fand three : other jfira^ar examples, 
&p. fflmprSed in the following Table. ' . " 


[ * 2 ] 



I>3 3 

• Maxim 4th. The. apir fun Being the fame, the effe3 
witl'Ae mttrly as the -cube of the vekcityoftbe water. 

This alfo will appear by comparing the contents 
of columns 3, 8, and 10 ; as for 

Example 1/?, o/ : N\i,< and N°. 10, viz. 

N°. Turns, ' Expence. Effedh 

\ 1 ""„ " ■ ■ '"88 "r^ -27f — -— 141 1 

10 ■ ■ ■ 42 — *- 114 — 117 

Lemma. It muft here he dbferved, that if water 
paffes out of ab apdrture, in the fame fe&ioti, but 
with different > velocities $ the expence will bp pro- 
portional to the velocity; and therefore convierfely," 
if the expence is not proportional to the velocity, 
the fedtiori of the water is not the fame. 

Now comparing the water difcharged with the 
turns of N?. i. and 10, we fhall have 88 :54a :: 
275 : 1 3 1, a ; but the water difcharged by N°. 10. 
is only 1 14 lb. therefore, tho' the fluice was drawn 
to the feme height in N ^ 10. *s in N°. 1. yet the 
fe&ion of the water palling out, wai.lefs in N°. 10. 
thanN°> i; in ihe proportion of 1*4 to 131,2 $ con- 
fequently had the effective aperture or fe&ion of the 
water been the fame inN c . 10. as in N°. 1. fothat 
131,21b. of water had been difcharged irrftead of 
114, the effedfc would have been increafed in. the 
fame proportion 4 that is, 

!by the Lemma, 88 : 42 : : 277 : 131,2 
by maxim ift, 114 : 131,2 : : 117 : 134,5 

andty max. 4 th , [ 68 ^;. 7 $b} s: I4M: J2H. 

Difference — — ip 


. [Hi ■ 

Ute 4Fe& ttwtftfcre bf N°> ra compared #ith 
N°i i. is Ifefc than it bughi to be by to ptfMft 
maxim in the rati© of 7 : 8. 

The foregoing, aiifi Inree other ftmfoar examples,, 
are cbMai^a in me fcllotoihfr Tkble. 


-*A H?UOI* . 
: -JodfOiJ 





. 8 

4 T**^ c^ 
I & a? 3 \ «ft $ 

* ft *A ; 


Sv4 ' 

C* ►* ^fr 

•• •• 1. 

«> c« c* 

u - •• 



•• •• •• 

00 « £ °° ^oo 



*• •« 

••■ •* 

fVt "* I" 

i. :' ». I ••• *• ••' ' 


^S ** 

• • . • 

CO CO * 

•Y V •• 

•• •• ^. 



• • •» 


10 ¥h- I VO.^ 

M H* f O 00 

T '« 

„ iv«06 ' 

Ji!- : 

/ajnuwu;* . *,* i 

v " in sftjnj. 

I l'<Hi&i : 

00 ^ 

1^ ^f 






Os ] 


Ob/erv. ift. Qn comparing column ad and 4th, 
Tab, I. it is evident, that the virtual bead bears 
ao certain proportion to the bead of water $ but that 
when the aperture is greater, or the velocity of the 
water iffuing therefrom lefe, they approach nearer to 
a coincidence : and confequently in the large open- 
ings of mills and flukes, where great quantities of 
water are difcharged from moderate heads, the head 
of water, and virtual head determined from the ve- 
locity, will nearly agree, as experience confirms. 

Obferv. 2d. Upon comparing the feveral pro* 
fxations between the pmier and effeB in column 
nth, the moft general is that of 10 to 3 $ the ex- 
tremes 10 to 3,2 and 10 to 2,8 5 but as it is obferv-* 
able, that where the .quantity of water, or the ve- 
locity thereof; that is, where the power is greateft, 
the 2d term of the ratio is greateft alfo : we may 
.therefore well allow the proportion fubfiftiog in large 
works, as 3*0 1. 

Obferv. 3*/. The proportions of velocities be- 
tween the water and wheel in column 1 2, are con- 
tained in the limits of 3 fa 1 and 2 to ij but as 
the greater velocities approach tfie limit of 3 to i # 
and the greater quantity of water approach to that 
of 2 to 1, the beft general proportion will be that 
of 5 to a. 

Ob fern. 4A&. On comparing the numbers 'in 
column *3» it appears, that there is no certain ratio 
.between the .load that the wheel will carry at its 
.«w,x/»w, and what will totally flop it; bijt tji^t 
.they are contained within the limits or 2Q to 10, and 

of 20 to i£ 5 but as the effe& approaches neareft 
to the ratio of 20 to 15, or of 4 to 3, when the 
power is greateft, whether by increafe of velocity, or 
quantity of water, this feems t9 be the moft appli- 
cable to large works : but as the load that a wheel 
ought to have, in order to work to the heft advan- 
tage, can be affigned, by knowing the effect it ought 
to produce, and the velocity it ought to have in pro- 
ducing it ; the exa<St knowlege of the greateft load it 
will bear, is of the lefs confequence in practice. 

It is to be rioted, that in all the examples under the 
three laft of the four preceding maxims, the effect 
of the lefler power falls ihort of its due proportion 
to the greater, when compared by its maxim ; except 
the laft example of maxim 4th : and hence, if the 
experiments are taken ftriftly, we muft infer, that 
the effects increafe and diminifh in an higher ratio 
than thofe maxims fuppofe : but as the deviation is 
not very confiderable, the greateft being about i*8th 
of the quantity in queftion ; and as it is not eafy to 
make experiments of fo compounded a nature with 
abfolute precifion ; we may rather fuppofe, that the 
lefler power is attended with fome friction, or works 
under fome difadvantage, which has not been duly 
accounted for, and therefore we may conclude, that, 
thefe maxims will hold very nearly, when applied to 
works in large. 

After the experiments above mentioned were tried, 
the wheel, which had originally 24 floats, was re- 
duced to twelve $ which caufed a diminution in the 
efleft, on account of a greater quantity of water 
cfcaping between the floats and the floor j but a cir- 


[ *7 3 

cular fweep being adapted thereto, of fuch a length, 
that one floit entered the curve before the preceding 
one quitted it, the effedt came fo near to the former, 
as not to give hopes of advancing it by increafing 
the number of floats beyond 24 in this particular 

P A R T II. 

Concerning Overshot Wheels, 

Read May 24, JN the former part of this eflay, we have 
- I 7S9- J[ confidered the impulfe of a confined 
jftream, adting on Under/hot Wheels. We now pro- 
ceed to examine the power and application of water, 
when adting by its gravity on Overjhot Wheels. 
, In reafoning without experiment, one might be 
led to imagine, that however different the mode of 
application is ; yet that whenever the fame quantity 
of water defcends thro' the fame perpendicular fpace, 
that the natural effedtive power would be equal; 
fuppofing, the machinery free from fridtion, equally 
calculated to receive the full effedt of the power, and 
to make the moft of it : for if we fuppofe the height 
of a column of water to be 30 inches, and refting 
upon a bafe or aperture of one inch fquare ; every 
cubic inch of water that departs therefrom will ac- 
quire the fame velocity or momentum,, from the 
uniform preffure of 30 cubic inches above it, that 
one cubic inch let fall from the top will acquire in 
/ailing down to the level of the aperture ; viz. fuch 
a velocity as in a contrary direction would carry it to 

Da the 

f «1 

the level from whence it fell ; * one would therefore 
fuppofe, that a cubic inch of water, let fail thro* a 
fpace of 30 inches, and there impinging upon an- 
other body, would be capable of producing an equal 
effect by collifion, as if the fame cubic inch had de- 
fcended thro' the fame fpace with a flower motion,, 
and produced its effedts gradually : for in both cafes 
gravity adts upon an equal quantity of matter, thro* 
an equal fpace -f*; and confequently, that whatever 
was the ratio between the power and efFedt in under- 
fhot wheels, the fame would obtain in overfhot, and 
indeed in all others : yet, however conclufive this 
reafoning may feem, it will appear, in the courfe of 
the following deductions, that the efiedt of the gra- 
vity of defcending bodies is very different from the. 
effedl of the ftroke of fuch as are. hon-elafiic> tho* 
generated by an equal mechanical power. 

The alterations in the machinery already defGrrbedi. 
to accommodate the fame for experiments on over*- 
fhot wheels, were principally as follows^ 

Plate V. Fig. z* The fluice lb being (hut 
down, the rod HI was unfcrewed and taken off. 

The underftiot water-wheel was takeiuofF the axis,, 
and inflead thereof an overfhot wheel of the fame. 

* This is a confequence of the rifing of jctts to the height of' 
their refervoirs nearly. 

f Gravity, it is true, afls a longer fpace of time upon the body 
that defcends flow than upon that which falls quick ; but this can- 
not occafion the difference in the effed : for an elaftic body falling- 
thro* the fame fpace in the fame time, will, by collifion upon an- 
other elaftic body, rebound nearly to the height from which it fell 5, 
or, by communicating its motion, caufe an equal one to afcend 
to the fame height. 



diariieter was put into its place- Note, This wheel 
was two inches in the fhroud or depth of the bucket $ 
the number of the buckets was 30 . 

The ftandards S and T, Fig. 1. were raifed half 
an inch, fo that the bottom of the wheel might be 
clear of ftagnant water. 

A trunk, for bringing the water upon the wheel r 
was fixed according to the dotted lines f g y Fig. 2. 
The aperture was adjufted by a {huttle b i y which 
alfo clofed up the outer end of the trunk, when the 
water was to be -ftoppecL 

Fig. 3. The ratchet v y riot being of one piece 
of metal with the ferrule e *, i i (tho' fo defcribed 
before, to prevent unneceffary diftindfions), was with, 
its catch turned the contrary fide } consequently the 
moveable barrel would do its office equally, notwith- 
ftanding the water- wheel, when at work, moved the 
contrary way- 


Specimen of a Sett of Experiments. 

Head 6 inches. 
a 4 \ ftrokes of the pump in a minute, ia ditto =s 

80 ib. * 
Weight of the fcale (being wet) io£ oz. 
Counterweight for ao turns, befides the fcale, 3 oz. 

Weight in 
No. the Scale. Turns. Produ&. Obfervations. 

1 — — olb. — 60 — — 1 Threw moft part 

2 - ■ ■ ■ . 1 ■■ 56 ■ ■■ ■ ■ * > of the water out 
2 . , , 2 ■ 52 ■ ■ ■ 1 of the wheel. 

4 3 — 49 — 1+7 1 Received the wa- 

5 ■ 4 — — 47 ■ 188 J cer more quietly. 

6 5 45 — 225 

7 6 42 i 255 

8 — — 7 — — 41 — 287 

9 ,. 8 38* 308 

10 9 36! 328* 

11 10 35 i 355 

12 — 11 — — 32 £ — — 360^ 

13 12 31 i 375 

- 14 13 28| 370 j- 

15 14 27 i 385 

16 15 26 ■ 390 

17 16 24 1 392 

18 — — 17 224 — — 3864 

19 18 -— 21 4 391 1 

20 19 20 | 394 }1 Maximum. 

21 — 20 19 i 395 J 

22 21 184 388i 

23 ■ 22 ■ 18 ■■ 396 Work'd irregular. 

24 23 — - - Oirerfet by its load. 

* The fmall difference, in the value of 12 ftrokes of the pump, 
from the former experiments, was owing to a fmall difference in 
the length of the ftroke, occafioned by the warping of the wood. 


[ 5» ) 

Reduction of the preceding Specimen* 

In thefe experiments the head being 6 inches, and 
the height of the wheel 24 inches, the whole de- 
fcent will be 30 inches : the expence of water was 
14 i ftrokes of the pump in a minute, whereof 12 
contained 80 lb.; therefore the water expended in a 
minute was 967 lb. which, multiplied by 30 inches, 
gives the power = 2900. 

If we take the 2 oth experiment for the maximum, 
we fhall have 20 f turns in a minute, each of which 
raifed the weight 4! inches, that is, 93,37 inches in 
a minute. The weight in the fcale was 19 lb, the 
weight of the fcale io-r oz. ; the counter-weight 3 oz. 
in the fcale, which, with the weight of the fcale 
10 £ oz. makes in the whole ao-lb. which is the 
whole refiftance or load : this, multiplied by 93,37 
inches, makes 1914 for the efFedt. 

The ratio therefore of the power and effeB will 
be as 2900 : 1914, or as 10 : 6,6, or as 3 : 2 nearly. 

But if we compute the power from the height <5f 
the wheel only, we fhall have 96 •§ lb. multiplied by 
24 inches = 2320 for the power > and this will be 
to the effeSi as 2320 ; 19*4, or as 10 : 82, or a* 
5 : 4 nearly. 

. The reduction of this fpecimen is fet down in 
N°. 9. of the following Table $ and the reft were die- 
dufted from a fimilar fett of experiments, reduced ia 
the fame manner. 




1 3* 3 

Table IL containing the Refujt of Sixteen Setts 
of Experiments on- Over/hot Wheels. 

■•** ■* 











. 9 





















6 |f 




" s 

is -2 


9 6f 


20 j 









'• 2 - 







20 J 


5 ** 
~ g 

£ O 

•1 53014360 





2 3i 



is .. 














.2? 20 





I -9l 
21 i 




1 61 










10: 6,9 
iq : 6,9 
1 o : 7, 6 
10 : 7,3 
10 : 7,3 


o Jq - 

.2 "tj 

(2 °- 



.10 : 6,8 

10 : 6,$ 
10: 6,6 


10 : 7fi 
jo :6,4 
10 : 8,2 
10: 8,2 




10 : 8,4 
10 : 8, 






10 : 6,1 
10 : 5,9 

10 :8,i 
10: 8,2 
10-: 8,2 






10 : 6,5 
10; 5,2 

to : 8,4 
ro : 8,1 


10: $ 9 6 

10: 7,6 

9. I 10. 




Obfervations and r £)edu£iions from the foregoing Ex* 


I. Concerning the Ratio between the Power and Effift 

of Over/hot Wheels. 

The effedtive power of the water muft be rec- 
koned upon the whole defcent j becanfe it muft be 

4 raifed 


*aifcd tfett height* in Order to be in a condition of 
producing the ftme effect a fecond time. 

The ratio's between the powers fo eftimated, and 
the effe&s at the maximum deduced from the feveral 
ftm of experiments, we exhibited at one view in 
column 9. of Table H. ; and from hence it appears, 
that thafe ratio's differ from that of 10 to 7,6 to that 
of k>: ft*, that is, nearly from 4: 3 to 4: 2. In 
thofe {experiments where the heads of water and 
quantities expended are foaft* the proportion kneasv 
ly 4s 4 : 3 j but where the heads aad quantities are 
greateft, it approaches nearer to that of 4 : 2 ; and 
by a medium of the whole, the ratio is that of 3 : z 
nearly. We have feen before* in our obfervations 
upon the effects of underfoot wheels, that the gene- 
rail ratio of the power to the effe£, when greatefl* 
was 3:1; the effeSl therefore tfoverjhot wheels \ under 
the Jhme circumjtances of quantity and fall, is at a 
medium double to that of the underfoot : and, as a 
confequence thereof, that nonelaftic bodies^ when a£l- 
ing by their impuffe or coffijkn, fohtmtmicate oniy a 
fart of their original power \ the other part beifcg 
ibent in changing their figure in confequence of the 

The powers of water computed from the height 
of the wheel oniy, compared with the effefts, as in 
cohirnn 1 o. appear to obfcrve a more conftant ratio : 
for if we take die medium of each clafs, which is 
fef down in colnmn 1 1 % we fliall find the extremes 
to dfflfer no more than from the ratio of 10 : 8,1 to 
that tif to : 8,5 5 and avthe fecond term of the ratio 
«*r&d&atty increafes fromS,t toS,f, by an increale 
rf head from 3 inches to ri, tfae excefs of 8, J above 

E 5,i 



[ 34 I 

8,1 is to be imputed to the fuperior impulfe of the 
water at the head of n inches above that of £ 
inches : fo that if we reduce 8,i to 8, on account 
of the impulfe of the 3 inch head, we jhall have 
the ratio of the power, computed upon the height of 
the wheel only, to the effeSi at a maximum as 10 :S, 
or as 5 : 4 nearly : and from the equality of the ratio 
between power and efFedt, fubfifting where the con- 
ftrudtions are fimilar, we muft infer, that the effefts, as 
well as the powers, are as the quantities of water and 
perpendicular heights multiplied together rejfye&ively \ 

II. Concerning the mofl proper Height of the Wheel 
in proportion to the whole Defcent. 

We have already feen, from the preceding ob~ 
fervation, that the eflfecft of the fame quantity of 
water, defcending thro' the fame perpendicular fpace, 
is double, when adting by its gravity upon an over^ 
.(hot wheel, to what the fame produces when adting 
by its impulfe upon an underfhot. . It alfo appears, 
that by increafing the head from 3 inches to 1 1, that 
is, the whole defcent, from 27 inches to 35, or in the 
ratio of 7 to 9 nearly, the effedt is advanced no more 
than in the ratio of 8,1 to 8,4, that is > as 7 : 7,26; 
and confequently the increafe of efFedt as not i-7th 
of the increafe of perpendicidar height Hence it 
follows, that the higher the wheel is in proportion to 
the whole defcent 7 the greater will be the effeft ; be*- 
caufe it depends lefs upon the impulfe of the head, 
and more upon the gravity of the water in the 
buckets: and if we confider how obliquely the 
water iffuing from the head muft ftrike the buckets^ 
we (hall not be at a loft to account for the little ad- 


vantage that arifes from the itapulfe thereof ; and fhali 
immediately fee of how little cohfequence this impulfe 
ds to the efle£t of an dverfhot wheel. However, as 
.every thing has its limits, fo has this : for thus much is 
xlefirable, that the water Jhould have fomewhat greater 
velocity > than the circumference of the wheel, in 
coming thereon •, otherwife the wheel will not only 
<be retarded, by the buckets ftriking the 1 water; but 
thereby dafliing a part of it over, fo much of the 
power is loft. 

The velocity that the circumference of the wheel 
ought to have, being known by the following de- 
du&ions, the head requifite to give the water its pro- 
per velocity is eafily computed from the common 
rules of hydroftatics $ and will be found much lefs 
than what is generally prattifed. 

i< ■ ~ 

III. Concerning the Velocity of the Circumference of 

the Wheel \ in order to produce the greateft EffeB. 

If a body is let fall freely from the furface of the 
head to the bottom of the defcent, it will take a 
certain time in falling ; and in this cafe the whole 
a&ion of grayity is fpent in giving the body a certain 
velocity : but if this body in falling is made to aft 
upon fome other body, fo as to produce a mechani- 
cal effedt, the falling body will be retarded ; becaufe 
;apart of the adtion of gravity is then fpent in pro- 
ducing the effedt, . and' the remainder . only giving 
motion tQ ihe falling body : and therefore toe flower 
ja body deft ends, the greater will be the portion of the 
aftion of gravity applicable . to , the producing a me- 
chanical jffeft \ and in confequence the greater that 
"pfFeft may be; ^ ^ 

E 2 U 

If a flrtam of water falU ioto tfte bucket of an 
.over&ot wheel, it is there retained till the wheel 
by moviog round difeharges it : of confequeace the 
flower the wheel moves, the more water each bucket 
•will receive : fo that what is loft m fpeed* is gained 
by the preflure of a greater quantity of Water a&tng 
in the buckets at once : and, if confidered ody in this 
light, the mechanical power of m ovcrfhot wheel to 
jjroduce effects will be equal, whether it moves quick 
or flow : but if we attend to what has been juft now 
obfervod of the felling body, it wifl appear that 
fo much of the ax^an of gravity, as j» empfoyai ia 
giving the wheel and water iberon a greater velocity* 
couft bef ubtraded from its pcefiWre upon the buckets 5 
fo that, ttho' the prated: made by mukiplying the 
number of cubic inches tof water a&iisjg; in the what 
at once by its velocity will be the fame in all cafes * 
yet, as each eubfc inch, when the metecky is greater 
does not pre& fo much upon the bucket as when it 
is left, the power of the water to produce effects will 
be greater in the lefs velocity than in the greater z 
and hence we are led to tnis general rule, that r 
caeteris paribus, the lefs the velocity rf. the wheel, the* 
greater will be the <effe& therepf. A confirmation of 
this doftrine, together with the limits it is fubjeft to 
in praftice, may be deduced from the foregoing fpe- 
cimen of a fett of experiments. 

Fromthefe experiments it.appears,. that when the: 
wheel made about 20 turns in a minute, the effedt: 
was, near upon, the greateft. When it made 30 turns, 
the effect was diminifhed about -jV part $ but that 
when it made 40, it was dimimifhed about £; when* 

it made left than 18 i, its motion was irregular ; and 


r 37 j • 

when ft was loaded fo as not to admit its making i $ 
turns, the wheel was overpowered by its load. 

It is an advantage in praftice, that the velocity of 
the wheel fhould not be diminifhed further than what 
will procure fome folid advantage in point of power ;. 
becaufe, catfris paribus, as the motion is flower* 
the buckets ipuft be made Urger; and the wheel 
being more loaded with water, the ftrefs upon ever^ 
part of the work will be increafed in proportion : 
The befi velocity for pra&tice therefore will befucb y at 
when the wheel here ufed made about 30 turns in a 
minute \ that is, when the velocity of the circum- 
ference is a little more than 3 feet in a fecond. 

Experience confirms, that this velocity of 3 feet 
in a fecond is applicable to the higheft overfhot wheels,, 
srs well as the loweft 5 and all other parts of the 
work being properly adapted thereto, will produce 
very nearly the greatefjfc *f&£t poffihle : however this 
alfo is certain from expedience, that high wheels may 
deviate further from this rule, before they mil lope 
their poiper r by a given aliquot part of the whole y 
than low ones can he admitted to do y for a wheel of 
64 feet high may move at the rate of fix feet per 
fecond without lofipg any conlktacablc part of its 
fiower * y and, on toe other hand, I have feen a 
wheel of 33 feet high, that has moved very fteadily 
and well with a velocity but little exceeding a feet*. 

* The 24 feet wheel gping at 6 feet in a fecond feems owing to 
Jbe final! proportion that the head (ueqiiifite to give the water the 

f^oper velocity #f jfe whsel) bean to d* whole height. 

IVY Con- 

• • * 

IV. Concerning the Load for ; an, Over/hot Wheel^ in 
' ,: ; order, that it may produce a Maximum. \ 

The maximum Joad for an over/hot wheel, is that 
-which reduces the circumferences of the wheel to its 
proper velocity - y and this will be known, by dividing 
^the effe£t it ought to produce in a given time by the 
fpace intended to be'defcribed by the circumferefrce 
of the wheel in the fame time : the quotient will be 
the refiftance overcome at the circumference of the 
wheel ; and is equal to the load required, the fric- 
tion and refiftance of the machinery included. 

V. Concerning the great eft pojfible Velocity of an . 

Over -/hot Wheel. 

:, The greateft velocity that the circumference of an 
<overftiot wheel is capable of, depends jointly upon 
*he diameter or height of the wheel, arid the velo- 
city of falling bodies $ for it is plain that the velocity 
of the circumference can never be greater, than to 
defcribe a femi-pircumference, while a body let fall 
from the top of the wheel will defcend thro' its di- 
ameter^ nor indeed quite fo great, as a body de- 
scending thro' the fame perpendicular fpace cannot 
perform the <fame in fo fmall a time when pafiihg 
thro* a femi-cirole, as would be done in a perpendi- 
cular line. Thus, if . a wheel is 1 6 feet i inch high, 
a~body will fall thro' the diameter in one fecond: 
jthis whed therefore can never af rive at a velocity 
equal to the making one turn in two feconds ; but, 
in reality, an overftiot wheel can never. come near 
this velocity; for when it acquires a certain fpeed, 
j the 

E 39 3 

ttie greateft part of the water is prevented" from en- 
tering the buckets $ and the reft, at a certain point 
of its defcent, is thrown out again by the centrifugal 
force. This appears to have been the cafe in the 
three firftf experiments of the foregoing fpecimen ; 
but as the velocity, when this begins to happen, de- 
pends upon the form of the buckets, as well as other 
circumftances, tbe^utmoft velocity of over/hot wheel* 
is not to be determined generally : and, indeed, it is 
the lefs neceffary in pra&ice, as it is in this circum- 
ftance incapable of producing any mechanical effect ^ 
for reafons already given. 

VI. Concerning the greatejl Load that an Over/hot 

Wheel can overcome. 


The greatejl load an over {hot wheel will overcome ^ 
eonfidered abfiradiedly, is unlimited or infinite : for as 
the buckets may be of ahy given capacity, the more 
the wheel is loaded, the flower it turns / but the 
flower it turns, the more will the buckets be filled 
with water ; and confequently tho' the diameter of 
the wheel, and quantity of water expended, are both 
limited, yet no refiftance can be afligned, which it is 
not able to- overcome : but in practice we always 
meet with fomething that prevents our getting inro 
infinitefimals 5 for when we really go to work to build 
a wheel, the buckets muft neceflarily be of fomc 
given capacity; and confequently fuch a rejijlance 
willjlop the wheels as is equal to the effort of all the 
buckets in one femi-circumference filled with wafer. 

The ftru&ure of the buckets being given, the 
quantity of this effort may be afligned $ but is not 
of much confequence to the practice, as in this cafe 


[ 4*> ) 

alfo the vrheel lofts its power j for tho 9 herd is thft 
exertion of gravity upon a given quantity of water, 
yet being prevented by a counterbalance frocq mov- 
ing, is capable of producing no mechanical ejfie&i 
according to our definition. But, in reality* an over-* 
{hot wheel ^generally ceafes to be ufeful before it if 
loaded to that pitch; for when it mtets witbjkcha 
rtfftance as to diminijh its velocity H a certain degree^ 
its motion becomes irregular $ yet this never happens 
till the velocity of the circumference is lefs than zjeet 
ferfecond > inhere the rejiftanct *j tquakle, as appear* 
not only from the preceding fpecimen, but from ex- 
periments on larger wheels. 


Havkvg now examined the different effeds of the 
power of water, when acting by its impulfe^ and by 
its weighty under the titles of underjhot and overjbtt 
wheels ; we might naturally proceed to examine the 
effe&s when the imputfe and weight are combined, as 
in the feveral kinds of br*aft-wbeels % &c. but, what has 
been already delivered being carefully attended to, the 
application of the fame principles in thefe mixt cafes 
will be eafy, and reduce what I have to &y on this 
head into a narrow compafs : for all kinds of wheels 
where the water cannot descend thro' a given fpace § 
imkfs the wheel moves therewith, are to be con* 
fidered of the natttre of an overihot wheel, accord- 
ing to the perpendicular height that the water de~ 
fcends froto ; and all thofe that receive the impuMb 
4&r (hock of die water > whether in an horizontal* per- 
pendicular, or oblique direction* Are to be considered 
B& understate Atd thetfefer e a wk*l> which the 



water ftrikes at a certain point below the furface of 
the head, and after that defcends in the arch of a 
circle, preffing by its gravity upon the wheel ; the 
effeB offuch a wheel will be equal to the effeSl of an 
under/hot, whofe bead is equal to the difference of 
level between the furface oj the water in the refervoir 
and the point where itjlrikes the wheel, added to that 
of an over/hot, whofe height is equal to the difference 
of kvel> between the point where it Jlrikes the wheel 
and the level of the tail-water. It is here fuppofed, 
that the wheel receives the (hock of the water at 
right angles to its radii ; and that the velocity of its 
circumference is properly adapted to receive the ut- 
moft advantage of both thefe powers ; otherwife a 
reduction muft be made on that account. 

Many obvious and considerable improvements up- 
on the common pradtice naturally offer themfelves, 
from a due confideration of the principles here eftab- 
lifhed, as well as many popular errors fhow them- 
felves in view : but as my prefent purpofe extends 
no farther than the laying down fuch general rules 
as will be found to* anfwer in pra&ice, I leave the 
particular application to the intelligent artift, and to- 
the curious in. thefe matters, m 

P A R T lit 

On the ConJlruSlion and EffeSts c/WrNDMkL- 


Read 31 May &r | N trying experiments on windmill- 
14 June, 1759- 1 f ai i s> ^ wind itfelf is too uncertain 

to anfwer the purpofe : we muft therefore 
courfe to an artificial wind. 

F Tbi* 




- This. : njay [bp/4<Hie two. ways.; either by^ca^fiog 
the air to moye againft tlj? machine, or the raa$hif*r 
to move again^ the air. * To canfe the air to ih&ve 
^ainft the machine, ia a fufficiefit vplumn, with 
fteadinefs and. the requifit? velocity, is n$t eafily put 
in pra&ice: To parry the machine forward in a 
jright line again ft the air, wquH require: a larger room 
than I could conveniently meet with. What I found 
moft practicable, therefore, was,, to carry the axis, 
^whereon the fails were to b^fixsd, progreflivfely round. 
in the circumference of 3 large circle. Upon this 
idea * a machine wa$. conftrufted, . as follows* 

Plate VI. Fig. i. 
ABC is a pyramidical frame for - fuppprting the 
moving parts. 
D E is an upright axis, f whereon is framed 
F G, an arm for carrying the fails- at a proper dis- 
tance from the center of the upright axis. 
■ ■ ■ ■ i > 

* Some years ago Mr. Roufe, an ingenious gentleman of Har- 
borovgh in Letcefterikire, fet about trying experiments on the ve- 
locity of the wind, and force thereof upon plain, fucfaces and 
winarnill-faUs : and much about the fame time Mr, EUicott con- 
trived a machine for the ufe of the late* celebrated Mr. B, Robins, 
for trying the refiftance of plain furfaces moving thro' the air.- 
The machines of both thefe gentlemen were much alike, tho' ^t 
that time totally unacquainted with ^ach other's inquiries. But it 
often happens, that wheii two perfons think juftly upon th$ fame 
iubjed, their experiments are aJjke. This machine was alfo.built, 
upanJhe-fime idea is th$ foregoing ; but dirrered in having the hand 
for the firft mover, with a pendulum for its regulator, inftead of a 
weight, as in the former 5 which was certainly beft for the pur- 
ipofes ofmeafuring the impdfe of the windy or refiftance, b£ plain* : 
-but the latter is more applicable : to. experiments on windmill-fail* \ 
^becaufe every change^ of pofttion of tr*£\"ame fails will occafion 
their meeting the>ak wU^a4iffi^ent- Vekwaty, -tho- urged by the- 
jfejne weight. 

[43 J. 

If is a barrel Upon the upright axis, ~ whereon is 
wound a cord ; which* being drawn by the 
hand, gives a circular motion to the axis, and* 
to the arm F 6 ; and thereby carries the axis 
of the fails in the circumference of a circle, 
whofe radius is D.I, caufing thereby the fails 
: to ftrike the air, and turn round upon their 
own axis. 
At L is fixed the end of a fmall line, which paffing 
through the pullies M"N O, terminates upon a 
fmall cylinder or barrel upon the axis of the 
fails j and, by winding thereon, raifes* 
F the fcale, wherein the weights are placed for 
trying the power of the fails* This fcale, moving 
up and down in the direction of the upright 
axis, receives no^diflurbance from the circular 
<^R two parallel pillars iknding upon the arm FG, 
for the purpofe of fupporting and keeping 
fteady the fcale P$ which is kept from fwing- 
ing by means pf 
S T two fmall chains, which hang loofely round the, . 

two pillars. 
W is a weight, for bringing the center of gravity of 
the moveable part of the machine into the cen- 
ter of motion of the axis D E. 
VX is a pendulum* compofed of twp balls of lead,, 
which are moveable upon a wooden rod, and* 
thereby can be fo adjufted, as to vibrate in any 
time required. Thiy pe nd u lu m hangs upoft a : 
cylindrical wire, whereon it vibrates, as on a: 
rolling axis. 
Y is a perforated table for fupporting the axis of 
the pendulum, 

E z~ Nvte„ 

* \* 

[ 44 ] 

Note, The pendulum being fa adjufted, as to make 
two vibrations in the lime that the arm F G is in- 
tended to make one turn ; the pendulum being fet 
a vibrating, the experimenter pulls by the cord Z, 
with fufficient force to make each half revolution of 
the arm to correfpond with each vibration, as equal 
,as poffible, during the number of vibrations that 
the experiment is intended to be continued. A 
little practice renders it eafy to give motion thereto 
with all the regularity that is neceflary. 

Specimen of a Sett of Experiments. 

Radius of the fails — - — 2 1 inches 

Length of ditto in the cloth — — 18 

Breadth of ditto — 5,6 

% $ Angle at the extremity — — *— 10 degrees 

{, Ditto at the greateft inclination — 25 
20 turns of the fails raifed the weight 11, 3 inches 
Velocity of the center of the fails, in the! 

circumference of the great circle, in a>6f l . o in. 

fecond ■— — - — — — J 
Continuance of the experiment 52 feconds. 

N°. Wt. in the fcale. Turns. Produ<». 

I olb. 108 O 

2 — ^ 6 85 — - . 510 

3 6f 8* — fi6i 

^ _- y — y% -*, — £^(y 

$ 7*— 73 j+7f maxim* 

6 — r-* S *6f 520 


* In all the following experiments the angle of the fails is ac- 
counted from the plain of their motion ; that is, when they ftand 
at right angles to the axis, their angle is denoted o°, this notation 
being agreeable to the language of pra&itioners, >vho call the angle 
fo denoted, the weather of the fail j which they denominate greater 
or lefs, according to the quantity of this angle. 




N.B. The weight of the fcale and pulley was 3 pz.; 
and that 1 oz. fufpended upon one of the radii, at 
12 1 iches from the center of the axis, juft over- 
came the fridtion fcale and load of 7 i lb. 5 and 
placed at 14 £■§■ inches, overcame the fame refin- 
ances with <? lb. in the fcale. 

Reduction of the preceding Specimen. 

N°. 5. being taken for the maximum, the weight 
in the fcale was 7 lb. 8 oz. which, with the weight of 
the fcale and pulley 3 oz. makes 7 lb. 11 oz. equal to 
1 23 oz.; this added to the friftion of the machinery, 
the fum is the whole: refiftance *. The fri&ion of 
the machinery is thus deduced : Since 20 turns of 
the fails raifed the weight 1 1,3 inches, with a double 
line, the radius of the cylinder will be .18 of an 
inch -, but had the weight been raifed by a fingle 
line, the radius of the cylinder being half the former, 
viz. .09, the refiftance would have been the fame: 
we (hall therefore have this analogy; as half the 
radius of the cylinder, is to the length of the 
arm where the fmall weight was applied ; fo is the 
weight applied to the arm, to a fourth weight, which 
is equivalent to the fum of the whole refiftance to- 
gether; that is, .69 : 12,5 : : i oz. : 1390Z.: this 
exceeds 123 oz. the weight in the fcale, by 1 6 oz. or 
1 lb. which is equivalent to the fridtion ^ and which, 
added to the above weight of 7 lb. 11 oz. makes 
$ lb. 11 oz. = 8,6p lb. for the fum of die whole re- 

* The refiftance of the air is not taken into the account of 
refiftance, "bbcaufe it is infeparable from the application of the 

4 iiftance* 


fiffancc; and this,. multiplied by 7 j turns, makas * 
product of <Jj4, which may be called the representa- 
tive of the effeSt produced. 

In like manner, if the weight 9 lb. which caufpd 
the (ails to reft after being in motion, be augmented 
by die weight of the fcale and its relative fri&ion, it 
will become 10,371b. The refult of this fpecimen, 
is fet down in N°. 12. of Table III, and the refult 
of every other fett of experiments therein contained, 
were made and reduced in the fame manner. 


*& •* 

- C 47 3 

Table HI, Containing Nineteen Setts of Experiments onWindmilUSails 
of various Structures, Portions, and Quantities of Surfaces. 

The kind of fails 
made ufe of. 


Plain fails it att 
angle of 55°. 


P/*r>r fails weatherM C 
according to the< 
common pra&ice. C 

Weathered aceord-C 
ing to Macl4urin > s\ 
theorem. L 

Sails weathered in 
the Dutch man- 
ner, tried in va- 
rious pofitions* 

Sails weathered 1 in 
the Dutch manner, 
but enlarged to- 
wards the extremi- 

8 fails being fefio'rs f 
of elliffes in-theirx- 
beft portions. ( 

t r 1 


I 8 















18 J 





















* 2 f 


2 «: 















O 6* 

O g 

s «* 

r . *-» 

IT* cd 






7*5 6 










6 4 i 








1 2,09 




7»5 6 










J 4^7 8 













•* 5* B 

JJ > B 

8.8 1 


404 10:7 



f 404 






















io: 6,6 



to 15,8 



. o ** a 

" *« 2 

c<-2 £ 






ib : 8,4 




o ^ 



ib: 8,9 

ftp: 8,6 
J& : 8,4 

10: 10,1 
\o: 10,15 

10: 11,4 
10: 12,8 


10: 15,1 




. > 




Obfervations and Deductions from, the preceding 


I. Concerning the bejl Form and Pofition of Wind- 

In Table III. N°. i. is contained the refult of a 
fett of experiments upon fails fet at the angle which, 
the celebrated Monf. Parint, and fucceeding geome- 
tricians for many years, held to be the beft ; viz. 
thofe whofe planes make an angle ff° nearly with, 
the axis ; the complement whereof* or angle that the- 
plane of the fail makes with the plane of their mo 
tion, will therefore be 35 , as fet down in col. 2. and 
3. Now if we multiply their number of turns by 
the weight they lifted, when working to the greateft 
advantage, as fet down in columns 5. and 6. and* 
compare this product (col. 8.) with the other pro- 
ducts contained in the fame column, inftead of being 
the greateft, it turns out the leaft of all the reft. 
But if we fet the angle of the fame planes at fome- 
what lefs than half the former, or at any angle from* 
1 5® to 1 8°, as in N°. 3. and 4. that is, from 72 to 
7f° with the axis, the produdt will be increafed in. 
the ratio of 3 1 : 45* -, and this is the angle moft com*- 
monly made ufe of by practitioners, when the fun- 
faces of the fails are planes. 

If nothing more was intended than to determine the 
moft efficacious angle ta make a mill acquire motion 
from a ftate of reft, or to prevent it from pafling in to- 
reft from a ftate of motion, we fhall find the pofition 
of N°. 1. the beft ; for if we confult col. 7. which 
contains the leaft weights, that would make the fails 
pafs from motion tQ re A, we fhall find that of N°. 1. 


[ 49 ] 

(relative to the quantity of cloth) the greateft of all. 
But if the fails are intended, with given dimenfions, 
to produce the greateft effedt poftible in a given time, 
we muft intirely reject thofe of N°. i. and, if we 
are confined to the ufe of planes , conform ourfehes to 
Jome angle between N°. 3. and 4. that is, n$t lefs than 
7 **> or greater than 7 j> °, with the axis. 

The late celebrated Mr, Maclaurin has judicioufly 
distinguished between the action of the wind upon a 
fail at reft, and a fail in motion ; and, in confequence* 
as the motion is more rapid near the extremities than 
towards the center, that the angle of the different 
parts of the fail, as they recede from the center, 
ihould be Varied. For this purpofe he has furnifhed us 
with the following theorem *. c< Suppofe the velocity 
<f of the wind to be reprefented by a, and the velo- 
a city of any given part of the fail to be denoted by 
€i € ; then the effort of the wind upon that part of 
the fail will be greateft when the tangent of the 
an gle, in wh ich the wind ftrikes it, is to radius as 

« 1 2, 4. .iff _{- If to 1." This theorem then af- 

figns the law, by, which the angle is to be varied ac- 
cording to the velocity of each part of the fail to the 
wind : but as it is left undetermined what velocity 
any one given part of the fail ought to have in rcfpc^i 
to the wind, the angle that any one part of the fail 
ought to have, is left undetermined alfo ; fo that we 
are ftill at a lofs for the proper data to apply the theo- 
rem. However j being willing to avail myfelf thereof, 
and confidering that any angle from 1 5 to 1 8° was 
beft fuited to a plane, and of confequence the beft 

* Maclaurin's account of Sir Ifaac Newton's philofophical dif- 
covcries, p. 176, art. 29. 

G mean 


mean angle, I made the fail, at the middle diftancr 
between the center and the extremity, to ftand at an. 
angle of 1 5 41' with the plane of the motion 5 in 
which cafe the velocity of that part of the fail, whea 
loaded to a maximum, would be equal to that of 
the wind, or c » a. This being determined, the reft 
were inclined according to the theorem, as follows : 

Angle with Angle of 
the axis. weather. 

f£ - - c — \a - - 6f 26'-- 26* 34' 

Parts of theU--' = i*-- 6 9 54 - - 20 * 
radius from< f - - € =Z a - - 74 19 - - 15 41 middfe 
the center. * - - f = j|* - 77 20 - - 12 40 . 

i--* = aa--8i o - - 9 o extremity.. 

The reful t hereof was according to ft°. f. being 
nearly the fame as the plane fails, in their beft por- 
tion : but being turned round in their fockets, fo that 
every part of each fail flood at an angle of 3 , and 
afterwards of 6% greater than before, that is, their 
extremities being- moved from 9 to 12* and if , the 
products were advanced to ji$ and 527 refpedtively* 
Now from the fmall difference between thofe two> 
produ&s, we may conclude, that they Were nearly in 
their beft pofition, according to N°. 7. or fome angle 
between that and N 9 . 6 : but from thefe, as well as 
the plane fails and others, we may alfo conclude* 
that a variation in the angle of a degree or two. 
makes very little difference in the. effeSi y when the 
angle is near upon the bejl. 

It is to be obferved > that, a foil inclined by the 
preceding rule will expofe a convex furface to the 
wind : whereas the Dutch* and all our modern 


t 5* ] 

mill-builder*, tho^they make the angle to diminifiv 
in receding from the center towards the extremity, 
yet constantly do it in fuch manner, as that the far- 
face of the fail may be concave towards the wind. 
In this manner the fails made ufe of in N°. 8, 9/10, 
ii, 12, and 1 3. were conftru&ed ; the middle of the 
fail making an angle with the extreme bar of 1 z\ 
and the greateft angle (which was about -£• of the ra- 
dius from the centre) of 1 j° therewith, Thofe fails 
being tried in various portions, the beft appears to 
be that of N°. 11* where the extremities ftood at an 
angle of 7°! with the plane of motion, the produdfc 
being 639 : greater than that of thofe made by the 
theorem in the ratio of 9 : 1 1, and double to that of 
N°. i.j and this was the greateft product that could 
be procured without an augmentation of furface. 
Hence it appears, that when the wind falls upon a 
concave fur j ace > it is an advantage to the power of 
the whole 7 tho % every party taken feparately^Jhould not 
be difpofed to the beft advantage *. 

Having thus obtained the beft pofition of the fails, 
or manner of weathering, as it is called by workmen,; 
the next point was to try what advantage could be 


* By feveral trials in large I have found the following angles to 
anfwer as well as any. The radius is fuppofed to be divided into 
6 parts and i-6th, reckoning from the center, is called 1, the ex- 
tremity being denoted 6. 

Angle with Angle with the plane 

N°. the axis* of motion. 

1 72° — 1 8* 

2 71 ■ 19 

2 -, , 72 - - 18 middle 

4 , , 74 „.„ 16 

5 77i ■ ■ ia| 

6 ■■ 83 ■ ■ • ■ 7 extremity. 

G 2 made 

[5* J 

made by an addition of furface upon the fame ra- 
dius* For this purpofe, the fails made ufe of had 
the fame weather as thofe N°. 8. to 13, with an 
addition to the leading fide of each of a triangular 
cloth, whofe height was equal to the height of the 
fail, and whofe bafe was equal to half the breadth : 
of confequence the increafe of furface upon the 
whole was one fourth part, or as 4 : f. Thofe fails, 
by being turned round in their fockets, were tried in 
four different pofitions, fpecified in N°. 14, 15, 16, 
and 17; from whence it appears, that the beft was 
when every part of the fail made a greater angle by 
2°|, with the plane of the motion, than thofe with- 
out the addition, as appears by N°. 15. the product 
being 820 : this exceeds 639 more than in the ratio 
of 4 : 5, or that of the increafe of cloth. Hence 
it appears, that a broader fail requires a greater 
angle ; and that when the fail is broader at the ex- 
tremity v than near the center *, thisjhape is more ad- 
vantageous than that of a parallelogram *. 
f Many have imagined, that the more fail, the 
greater the advantage, and have therefore propofed 
to fill up the whole area : and by making each fail 
a fedtor of an ellipfis, according to Monfieur Paring 
to intercept the whole cylinder of wind, and thereby 
to produce the greateft effed poflible. 

«■■■ I' I ' ' ' I I ■ ■ f II ■ H I II I ■ 

* The figure and proportion of the enlarged fails, which I have 
found beft to anfwer in large, are reprefented in the figure, Plate VI. 
where the extreme bar is i-ld of the radius (or whip, as ft is called 
by the workmen), and is divided by the whip in the proportioa 
*>f 3 t° 5* The triangular or leading fail is covered' with board 
from the point downwards i-3d of its height, the reft with cloth 
as ufual. The angles of weather in the preceding note are beft 
for the enlarged fails aHb ; for in pra&ice it is found, that the fails 
|t£d better have too little than too much weather. 



We have therefore proceeded to inquire, how far 
the effect could be increafed by a further enlargement 
of the furface, upon the fame radius of which N Q . 
1 8 and 1 9 are fpecimens. The furfaces indeed were 
not made planes, and fet at an angle of 3 $°, as Parint 
propofed ; becaufe, from N°. 1. we learn, that this 
pofition has nothing to do, when we intend them to 
work to the greateft advantage. We therefore gave 
them fuch an angle as the preceding experiments in- 
dicated for fuch fort of fails, viz. 1 1° at the ex- 
tremity, and 22 for the greateft weather. By N°. 
18 we have the product 1079, greater than N°. 15. 
in the ratio of 7 : 9 3 but then the augmentation of 
cloth is almoft 7:12. By N°. 19. we have the pro- 
duct ii6f y that is greater than N p . if. as 7 : 105 
but the augmentation of cloth is nearly as 7 : 165 
confequently had the fame quantity of cloth -as in 
1^°. 18. been difpofed in a figure fimilar to that of 
N°. 15, inftead of the produtt 1059, we fhould 
have had the produdt 13865 and in N°. 19, inftead 
of the produ<ft 1 1 6f, we fhould have had a produd: 
of i860; as will be further made appear in the 
courfe of the following deductions. Hence it ap- 
pears, that beyond a certain degree, the more the 
area is crowded with fail, the lefs effect is produced 
in proportion to the furface : and by purfuing the 
experiments ftill further, I found, that tho' in N°. 
19. the furface of all the fails together were not 
more than 7~8ths of the circular area containing 
them, yet a further addition rather diminifhed than 
increafed the effeft. So that when the whole cylinder 
cf wind is intercepted, it does not then produce the 
greateft effeftfor want of proper interftices to efcape. 

[ 54 ] 

It is certainly defirable, that the (ails of windmills 
fliould be as lhort as poffible ; but at the fame time 
it is equally defirable, that the quantity of cloth 
fhould be the lead that may be, to avoid damage by 
fudden fqualls of wind. The beft ftru&ure, there- 
fore, for large mills, is that where the quantity of 
cloth is the greateft, in a given circle, that can be ; 
on this condition, that the efFed holds out in pro- 
portion to the quantity of cloth ; for otherwife the 
effect can be augmented in a given degree by a lefler 
increafe of cloth upon a larger radius, than would 
be required, if the cloth was increafed upon the fame 
radius. The mod ufeful figure therefore for practice, 
is that of N°. 9. or 10. as has been experienced upon 
feveral mills in large. 



a* ^* 


Si ** 




** ; u CM 

2 S 


3 c 









Ratio of the greateft 
load to the load at 
a maximum. 

co m 

* 0* 
oo 0\ 
• • • . 

o o 

1 1 

00 00 
• • »• 


Ratio of the grcateft 
velocity to the ve- 
locity at a maxim™. 


0k 0h 

vo vo 

• • »• 

O o 

1 1 

t^ CM 1 

vo vo 

• • •• 

o o 



Ratio of the two 



: iil 





1 o 

and greater velocity. 

1 ° 

1 oo 

| <o 

1 00 


1 OS 

1 tx 


Tarns of the fails 






Maximum load for 
the half velocity. 




1 "vr> 



vo CO 
Os Q 


O «0 

ci> cm 


o ^ 

co O 



Greateft load. 



^* vooo 

1 1 

oo co 

vo *• 


Load at the maxi- 

rx cm 



VO vo 

* * 

CO w 

o vo 

* * 



Turns of the fails at 


vo O 
VO co 


M Q 

VO m 


Turns of the (ails 

vo t>s 




^ oo 


Velocity of the wind 
in a fecond. 

,4 H* 


^- OS 


Angle at the extre- 

vo us. 

o o 

1-4 »N 



•1 CM 1 

1 co ^t- I 




II. Con- 

r s6 3 

II. Concerning the ratio between the vebtity of 
windmill fails unloaded \ and their velocity when 
loaded to a maximum. 

Thofe ratio's, as they turned out in experiments 
upon different kinds of fails, and with different in- 
clinations (the velocity of the wind being the fame) 
are contained in column i o of tab. III. where the 
extremes differ from the ratio of 10 : 7,7 to that of 
10 : 5,8 ; but the mojl general ratio of the whole will 
be nearly as $ : 2. Thus ratio alfo agrees fufficiently 
near with experiments where the velocity of the wind 
was different, as in thofe contained in tab. IV. col. 13. 
in which the ratio's differ from 10 : 6,9 to that of 
I o : f,5>. However, it appears in general, that where 
the power is greater, whether by an enlargement of 
furface, or a greater velocity of the wind, that the 
fecond term of the ratio is lefs. 

III. Concerning the ratio between the greatejl load 
that the fails will bear without flopping^ or what 
is nearly the fame things between the leaft load 
that will Jlop the fails, and the load at the maxi- 

Thofe ratio's for different kinds of fails and in- 
clinations, are colle&ed in col. 1 1. tab. III. where the 
extremes differ from the ratio of iq : 6 to that of 
10 : 9,2 ; but taking in thofe fetts of experiments 
only, where the fails refpedtively anfwered beft, the 
ratio's will be confined between that of 10:8 and of 
10:9; and at a medium about 10 : 8,3 or of 61 f. 
This ratio alfo agrees nearly with thofe in col. 1 4 of 
tab. IV. However it appears, upon the whole, that ■ 
in thofe inftances, where the angle of the fails or 



quantity^ of cloth were greateft, that the fccond term 

oFthte' ratio was lefs* 


IV. Concerning the effe&s of fails, according to 
the different velocity of the wind. 

Maxim i. The velocity of windmill fails > whe- 
ther unloaded^ or loaded Jo as to produce a maximum ^ 
is nearly as the velocity of the windy their Jbape 
and poftion being the fame. 

: This appears by comparing together the refpe&ive 
numbers of columns 4 and 5, tab. IV. wherein thofe 
of numbers 2, 4, and 6, ought to be double of num- 
bers 1, 3, and f: but as the deviation is no- where 
greater than what may be imputed to the inaccuracy 
of the experiments themfblves, and hold good exact- 
ly in numbers 3 and. 4 $ which fetts were deduced 
from the medium of a number of experiments, care- 
fully repeated the fame day, and on that account are 
C*6ft to be depended upon ; we may therefore con- 
clude the maxim true. 

Maxim 2. The load at the m&o&mum U nearly y 
but /bfnewbat lefs than, as the fquUre of the, vekcity 
of the windy the jhape and pofition of the fails be* 
ing the fame. - • - ' » : • .* 

This appears by comparing ' together the number? 
in c6L6. tab. IV. wherein thofe of/ ikihibers V4i 
and 6 (as" the velocity is double),' ought to* 1)6 qua- 
druple of thofe of numbers 1, 3, and 5 ; kiftfead of 
which they fallfhort, number 2. by y#, nflmber'4 
by tV>' and number 6 by -^ part >of' dfe .Whole. 
The greateft of thofe deviations is not more ■ conii- 
derable than might be imputed to the unavoidable 

H errors 

E 5« J 

wrore \n fiiftking thq experiments; bpt af ikofe 
experiments, as well as thofe of (lie grea£e# Iqa^ alk 
deviate the fame way ; and alfo coincide with fome 
txperime&ts - communicate to tne by Mr. Roufe 
upon the refifUnce of planes i I am led to fuppofe a* 
fmall deviation, whereby the load fails fhort of the 
fquares of the velocity* and. firicp the experiments 
N° 3 and 4. are moft to be depended upon, we 
muft -conclude, that when the velocity is double, the* 
load falls fhort of its due proportion by 7 r 7 , or, for 
the lake of a rosmd jamjobcr, by about ^ paf t of tfie 

" r • 

Maxim 3d- Tie effe&s of the f&nt faiU at a maxi- 
mum are nearly^ but /amewbat k/s than* as tkp 
eukei *f the velocity of the fwittd. 

It has already been proved,. Maxim, lfi> tftatr the 
velocity of fails at the maximum > is nearly^ as the ve~- 
locity of the wind ; f and by Maxim 2d, that the load 
at the maximum is nearly as the fquace of the fame 
velocity: if thofe two maximums woufd hold pre* 
rifely, it would be a< canfequence that the effedr 
would he in. a< triplicate ratio thereof: bow tbia 
agrees with experiment will appear by ^Gpiparing 
together the products in col. *» of tab, 4, wherein, 
thofe of V& z> 4* sod & .(jhc. velocity of the wind 
being double), ought to be o<9uple sf fchofc of N<> 1. 
Z* and 5. inftead of which they fall fhwt,,No 2. by $ 
W° 4, by T ^> aDC * No <J. by 3- part of the whole. 
Now, if we rely on N° 3 . and 4. as the turns of the 
fails are as the velocity of the wind ; and fince the 
load of the maximum falls fhort of the fquare of the 
yclocky by about ^part of the whole:, the product 


[ $9 ] 

*ntde bf %be mtfttipUGatkm <$f the tarns Irtto theload> 
fnuft alfo fall (hort of the triplicate ratio by about ^V 
part of the whole product. 

Maxim 4th. The had if the fame fails at the maxi- 
mum is nearly as the fqua^es, and their tfflSl as the 
tubes, of their humber of turns in a given time^ 

This maxim fiVay be efteertied a confequence of 
the three preceding ; for if die turns of the &Us are 
as the velocity of die wind, whatever quantities are 
in any given ratio o£ the velocity of the wind, will 
be in the fame given ratio of the ttjrns of the fails : 
and therefore, if the load at the tnaximum is as the 
fquare, or the efFcdfc as the cube, of the velocity of 
the wind* wanting -~5 P*t when the veldcity is 
-double ; the load at the maximum will alijo be -as the 
iquare, and the efFe& as the cube, of the number of 
turns of the fails in a given time, wanting in like 
manner ^ part when the number of turns are double 
in the £ame time. In the pfefent 4afe, if we com- 
pare the loads at the Maximum coi. 6. with the 
fquarts of the ntmafod: of tarns col, f. of N° i and .2. 
f and 6. or the products of the fame numbers col. 8. 
with the cubes of the number of tarns col. f. inftead 
of felling fhort, as N° $ and 4. they exceed thofe 
ratios : but as the fetts of experiments N° 1 and 1* 
5 and & are not to fee efteernfed of squil authority 
with thofe -rf N° 3 and 4. we touft not f ely upon 
them farther than to obferve, that in tvntyafing the 
ySF&fi *$*&* rf forg* fn*chine*i the direSt proportion 
'fif-theftyuares aird cubes ref$e8dvety % xbiU hid as near 
jOS the affefih thabjbfoes can be ofyervedt and there- 

H 2 fore 

fore be fufficient for pra&ical eftlmation, without any 

Maxim jth. When fails are loaded fo as to produce 
a maximum at a given velocity, and the velocity of 
the wind increafes, the load continuing the fame ; 
iftly, The increafe of effe£t r when the increafe of the 
velocity of the wind is fmall, will be nearly as the 
fquares ofthofe velocities: idly, When thevehciiy of 
the wind is double , the effects 'will be nearly as i o : 2 j% : 
But> $dly, When the velocities compared, are more 
than double of that where the given bad produces a 
maximum, the effeSls increafe nearly in a fmple ratio 
of the velocity of the wind. 

It has already been proved, maxim ift and 2& y 
that when the vfclocity of the wind is increafed, the 
turns of the fails will increafe in the fame proportion, 
even when oppofed by a load as the fquare of the ve- 
locity j fcnd therefore if wanting the oppofition of an 
increafe of load, as the fquare of the velocity, the 
turns of the fails will again be increafed in a fimple 
ratio of the velocity of the wind on that account alfo ; 
that is, the load continuing the fame, the turns of the 
fails in a given time will be as the fquare of the ve- 
locity of the wind ; and the effect, being in this cafe 
ds the turns of the fails, will be as the fquare of the 
velocity of the wind alfo ; but this muft be under- 
flood only of the firft increments of the velocity of 
the wind ; for, 

2dly, As the fails will never acquire above a given 
velocity in relation to the wind, tho' (he load was 
diminished to nothing \ when the load continues the 

fame } 

[6i ] 

feme, *the more the velocity of the wind* increafes 
(tho* the effect will continue to increafe) yet the 
more it will fall fhort of the fquare of the velocity of 
the wind ; fa that when the velocity of the wind is 
double, the increafe of effect, inftead of being as 
1 14, according to the fquares, it turns out as 10 : 27^, 
as thus appears. In tab. 4. coh 5). the loads of N° 2, 
4/ and 6. are the fame as the maximum lodds in 
col. 6. of N° 1, 3, and f. The number of turns of 
the fails with thofe loads, when the velocity of the 
wind is double, are fet down in col. 10. and the pro- 
ducts of their multiplication in col. 11 : thofe being 
compared with the products of N° 1, 3, and 5. coK 
8. furnifli the ratios fet down in col. 12. which at a 
medium (due regard being had to N° 3. and 4.) will 
be nearly as 10: 27-y. 3dly. The load continuing the 
fame, grows more and more ihconfiderable, refpe£t- 
ing the power of the wind, as it increafes in velocity; 
fo that the turns of the fails grow nearer and nearer a 
coincidence with their turns unloaded 3 that is, nearer 
and nearer to the fimple ratio of the velocity of the 
wind. " When the velocity of the wind is double^, 
the turns of the fails, when loaded to a maximum, 
will be double alfo ; but, unloaded, will be no more 
than triple, by dedudtlon 2d: and therefore the pro- 
duct could not have increafed beyond the ratio of 
10:30 (inftead of 10: 2 7-) even luppofing the fails 
not to have been retarded at all by carrying the maxi- 
mum load for the half velocity. Hence we fee, that 
when the velocity of the wind exceeds the double of 
that, where a conftant load produces a maximum,, 
that the increafe of effect, which follows the increafe 
of the velocity of the fails, will be nearly as the velor 



3 . _ 


tclty of the wind* and ultimately Iff that ratio pr&- 
rifely. ' Hence alfo We fee that Windmills, fuch sis 
thfe different fftefciei for raiflrig Water for dfakittg*, 
Afc. lofe faHich of their fell eflfe<a> Wh&i idtihg Sgaiiift 
one invariable opposition. 

W. Concerning the effe&s bf fails tf different magni^ 
tudes, tbt 'jfiru&Ure and p&fition being fimilar> add 
the velocity yfthe tvind fbejbne* 

Maxim 6. In fails of a fimUar figure and portion, 
the number of turns in a given tittle will be recipro- 
cally as the radius pr length of the fail. > 

The extreme bar having the fame inclination fb 
*he plain of its motion, and to the wind, j its velocity 
at a maximum will always be in a given ratio to the 
velocity bf the wind -, arid therefore, whatever be the 
radius, the abfolute velocity of the extremity of the 
fail will be the fame : arid this will hold good re- 
'fpe&ing any other bar, whofe inclination is the fame, 
;at a proportionable diftarice from the center ^ it there- 
fore follows, that the extremity of all fimilar fails, 
*with the lame /Wind, will have the fariie abfolute 
velocity 3 and therefore take a f^ace of time to per- 
form one revolution in proportion to die radius ; or, 
which is the fame thing, the number of revolutions 
in the fame given time, will be reciprocally as the 
length oftheiaii. 

Maxim 7. The had at a Maximum that fails of 
ajltnilar figure and portion ittill frOert&ne, at a givek 
M fiance from the center of motion ', *toill be afs the cuk 
wf the radius. 

4 Ceo- 


Geometry informs us, that in fimilar figures the 
ffyrfaces are as the fqqares of their fimilar fides ; of 
confeqjyence the quantity of cloth will be as the 
fi^are pf the radius : alfo in fimilar figures and pofi- 
t£op$, the impulfe of the wind,, upon every fimilar 
fe&ion of the cloth,, will be in proportion to the fur- 
face of that fedion; and confequently, the impulfe 
of the wind upon the whole, will be as the furface of 
the whole : but as die diftance of every fimilar fee- 
tion, from the center of motion, will be as the ra-> 
dius ; die diftance of the center of power of the 
whole, from the center of motion, win be as the ra- 
dius alfo ; that is, the lever by which the power a&s r 
will be as the radius : as therefore the impulfe of the 
wind, refpeding the quantity of cloth, is as the 
fquare of the radiu$, and the lever, by which it adts,. 
as the radius fimply ; it follows,, that the load which* 
the fails will overcome, at a given diftance from the: 
center, will be as the cube of the radius. 

/ Mmmfy.tfhrejfcdi qf fails of fimilar figure and 
$oJition y are as the fquare of the radius. 


By maxim 6. it is proved, that the number of re- 
volutions made in a givea time, are as the radius in- 
vecfely. Under maxim ji. it appears, that the length 
pf the lever, by which the power a<5ts, is as the radius 
diredlly ; therefore th$fe equal and oppofite ratios de- 
ftray oue another : but as in fimilar figures the quan- 
tity of cjoth is as the fquare of the radius, and the 
a<$tion of the wind is in propbrtipn to the quantity of 
sl^th, as alfo appears under qiaxim 7 ; it follows that 
the eftefit is as the fquare of the radius. 


« -.. 

[6 4 ] 

X^orol. i. Hence it follows, that augmenting Are 
3ength of the fail^ without augmenting the quantity 
of cloth, does is not increafe the power $ becaufe 
what is gained by the length of the lever, is loft by 
the flownefs of the rotation. . . 

Corol. 2. If fails are increafed in length, the 
breadth remaining the fame, thejefied will be as the 
radius. • . 

VI. Concerning the velocity of the extremities of 
windmill fails, in re/pe£l to the velocity of the 

Maxim p. The velocity of the, extremities of Dutch 

fails, as well as of the enlarged fails, in all their ufual 

pqfitions when unloaded, or even loaded to a maximum^ 

are confderably quicker than the velocity of the 


The Dutch fails unloaded, as in Tab. 3. N<> 8. 

made iao revolutions in fa": the diameter of the 
fails beiqg 3 feet 6 inches, the velocity of their ex- 
tremities will be $5,4 feet in a fecond ; but the velo- 
city of the wind producing it, being 6 feet in the 
fame time, we (hall have 6: 25**4: ;i 14,2 j in this 
cafe therefore, the velocity of their extremities was 
4,2 times greater than that of the wind. In like 
manner, the relative velocity of the wind, to the .ex- 
tremities of the fame fails, when loaded to a maxi- 
mum, making then 93 turns in 52", will be found to 
'be as x : 3,3 ; or 3,3 times quicker than that of the 

5 The 

The, following table contains 6 examples of Dutch 
fails, and 4 examples of the enlarged fails, indiffer- 
ent, portions, but with the conftant velocity, of the 
wind of 6 feet in a fecond, from table 3 : and alfo 
6 examples of Dutch fails in different portions, with 
different velocities of the wind, from table 4. 

Table V. containing the rath of the velocity of the 
extremities of windmill fails to the velocity tf tho 






I !3 





t *3 


I .' 

1 ■ 


J 7 











«5 . 


5 . 




'O.S s 
.2 '£ 8 

6 o 
6 o 

4 Q 

6 o 

6 o 

* UIJ'l, 

4 4* 
8 9 

4 4> 
8 9 

.4 4i 




11 Mil I III II 11 j ,1 uj, I 

H«PQ of we veloqty 
of the wind and ex- 
tremities of the fails. 

unloaded, loaded. 


r 66 J ... \ 

It appears from the preceding colle&ion of ex- 
amples, that when the extremities of the Dutch fails- 
are parallel to the plane of motion, or at right angles- 
to the wind, and to the axis, as they are made accord- 
ing to the common practice in England> that their 
velocity, unloaded, is above 4 times, and loaded ta 
a maximum^ above 3 times greater than that of the- 
wind : but that when the Dutch fails, or enlarged, 
fails, are in their beft pofitions, their velocity un- 
loaded is 4 times, and loaded to a maximum, at a. 
taedium the Dutch fails are 2,7, and the enlarged 
fails 2,6 times greater than the velocity- of the wind*. 
Hence we are nirnifhed with, a method of knowing 
the velocity of the wind, from obferving the velocity 
of the windmill fails %. for knowing the radius, and 1 
the number of turns in a minute, we fliall have the 
velocity of the extremities ; which, cliv^ded by the 
following divifors, will give the velocity \ of the. 

Dutch fails in their common pofitidnIf nl ^ ed +- a 

t loaded —3.3 

Dutch Oils in their beft pofition - ^f^tX^ 

Enlarged fails in their beft pofition {^J^ J 5 

From the above divifors there arifes the following; 
compendiums ; fuppofing the radius to be 30 feet,, 
which is, the moft ufual length in this country, and 
the. mill to be loaded to a maximum^ as is ufuallv the. 
cafe with corn mills ; for every 3 turns in a mtnute y , 
of the Dutch falls in their common pofition^ the wind 
will move at the rate of 2. miles an hour $ for every 
5 turns in a mtnutey of the Dutch fails in their heft 

" pofitim> 

[ 67 3 

fofition, the wind moves 4 miles an hour ; and for 
every 6 turns in a minute > of the enlarged fails in 
their beji pojition , the wind will move 5 miles. an 

The following table, which was communicated 
-to me by my friend Mr. Roufe, and which appears to 
have been conftrudted with great care, from a con- 
fiderable number of fads and experiments, and Which! 
having relation to the fubjedt of this article; I here 
infert it as he fent it to me ; but at the fame time muft 
obferve, that the evidence for .thofe numbers where 
the velocity of the wind exceeds jo miles an hour; 
do not feem of equal authority with thofe of 50 miles 
an hour and under. It is alfo to be obferved, that 
the numbers incol. 3. are calculated according to the 
fquare of the velocity of the wind, which, in mode- 
rate velocities, from what has been before obferved, 
will hold very nearly, t 

t . 

I z Tabl* 

* * <• 

% -4 

t 68 J 

« » . . 

" ' » ... * 

Table VI. containing the velocity ttnd force of 
windy according to their common appellations. 

Velocity x»f 

the Wind. 






5 ! 













8es ^ 

leu o 










o £ 



appdlstk>M> ofidwfow 


,020 _ 







9>9 6 3 

i7i7 x S 


it • 1 1 

J iilHi l m*m**~m+«4 

Hardly peftjfptiblc. 

, r Gentle, pleafant wind. 

ciPteafant brifk gale. 

J r » ** ... 

f Very brifk. 
1 High winds* 

I Very high. 

A florin or tempeft, 
A great ftorm. 
An hurricane. 

An hurircane that tears tip trees, carries 
buildings before it, &c. 

VII* concerning the ahfolute effedl, produced by a> 
given velocity of the windy upon fails of a given? 
magnitude and conftrufiion. 

It has been obferved by pra&itioners, that in mills- 
with 'Dutch fails in theccinmon pofition, that when 
they make about ij turns in a minute, they then 


> . . » 

. •#• 



-*rork itt a mean rate : that is, by the compendium* 
in the laft article, when the velocity of the wind is 

.$% miles an hour, or ia-|feetin a ieoond; vfrhich r 

:ia common phrafe, would be called zfrejb gale. 
The experiments fetdown in Tab. IV. N° 4, were 

; tried with a wind, whofe velocity was 81 feet in, a? 

dEscond; . confequently had thofe experiments been 

-tried with a wind, whofe velocity was i2«f feet in:a 
fttond, the cfifcft, hy maxim 3d, would have been 
3 times greater ; becaufe the cube of u\h 3 times 
greater than that of 8$. 

Emm Tab* IV. N04. ,we find, that the fails, v?h$ n % 
velocity of the wind was 8| feet in frfcconiy 
made rjo revolutions in a minute, with a load <?f 
*ZAS fc* >From tbe.raea&ces of .the, machine, pre- 
ceding the specimen taf&.&tt . of .experiments, wc 
find, Jthat 2 o revolutions pf the fails raifed, the fcafe 
and weight 1 u ,3 inches: 130 revolutions will there- 

. fore, nrife the- fcale y% >4f inches, -. which* multiplied 
fey *7>5 a fcr makes japrodtfft of i*&7> for the «f&^ 
oftbeDutdi fails in their heft pofition r that is, whoa 
the velocity of the wind is Si feet in a fecond : this 
produd therefore multiplied by ^ee,. will give 386r 
for the effe<£t of the fame fails, whence velocity $£ 
the wind is . 1.2? feet in a fecood. 

Defagiiliers makes the utmoft power of a m$n r 
when working fo as to he able to hold it for fopie 
hours, to be equal- to that of raifing an hogfhead of 
.water : to; feet h^hkuai minute. ;Nwv, * an .hog&ead* 
confifting of 63 ale gallons, .being redwed;inf<> 
pounds averdupois, and the height into inches ; the 
produd made by s multiplying thofe two/iwmbejrs 
will be 76800 j* whiqh is 19 itin^cs greater. than the 


I 7° '■] 

■fprodutt of the fails lafl>mentioned, at I'ffbet'ina 
•Second : therefore, by maxim 8th, if we multiply 
«the fquare root of 19, that is 4,46, by. 21 inches^ 
the-length of the fail producing the effe6t 3861, ws 
fhall have 93,66 inches, or 7 feet 94 inches for the 
radius of a Dutch fail in its beft pofition, whbfe mean 
-power (hall be equal to that of a man : but if they are 

-in their common pofition, their length muft be in- 

^reafed inthe ratio of the fquare root of 442 to that 

-of 639, as thus appears j 

! The ratio of the maximum produftsi of ffi°- 8 and 
«ii. Tab. IIL are as 442:639 ; but by maxim 8, 
the effects of fails of different radii are as the fquaife 
<if the radii ; confequently the the fquare roots of the 
products or effe&s, are as the radii fimply : ; and 

-therefore as the fquare root of 44a is to that of 
^39 > f° is 93,66 to 112 j66', or 9 feet 4| inches. . 
If the fails are of the enlarged kind, then from 
Tab. IIL N° 1 rand 15. we -fliall have the fquare 
*oot ef 8ao tovthatof 4539; : 93,66 : 82,8 inches, or 
6 feet 1 0$ inches : fo that in round numbers we fliall 
have the radius of a fail, of a fimilar figure to their 
refpedtive models, whofe mean power fhall be equal 

-to'tfhat of a man ; 

The Dutch fails in their common' pofition 9! feet. 
The Dutch /ails in their beft pofition — - 8 
The enlarged fails, ih their beft pofition — 7 

Suppofe now the radius of a* fail 30 feet, and 
4o be conftru^d * upon the model of the enlarged 
-fails, N° 14 or 15. Tab.'III. .diyiding 30 by 7<wje, 
HQiali have 4,28, the fquare of which is 18,3; and 
r £his, jaccotding to maxim 7, will be the relative 


[ 7« ] 

power of a foil of jo feet, to one of 7 feet ; that is* 
when working at a mean rate, the 30 feet fail will be 
equal' to the power of 18^3 men, or of 3-|-horfes ; 
reckoning 5 men to a. horfe : whereas the efFed of 
the common Dutch fails, of the fame length, being 
lefs in the proportion of 820:442,, will be fcarce 
equal to the power of 10 men,, or of 2 horfea. 

That thefe computations, are not merely fpecula- 
tive* but will nearly hold good when applied to 
works in large,. I have had an opportunity of verify- 
ing : for in a- mill with the enlarged fails of 30 feet, 
applied to the crufhingof rape feed, by means of 
two runners upon the edge, for making. oil; I oh— 
fcrved r that when the fails made n turns in a mi- 
nute, in which cafe the velocity of the wind was* 
about 13 feet in a fecond, according to article 6th, 
that the runners then made 7 turns in a minuter 
whereas 2 horfes, applied, to the fame ^ runners, 
fcarcely* worked them at the rate of 3.1 turns in, the 
fame time. Laftly,, with regard to the real fuperio- 
rity of the enlarged fails, above the Dutch fails as 
commonly made r it has fufficiently appeared, not 
only in thofe cafes where they have been applied to 
new mills, but where they have been fubuituted in 
the place of the others. 

yill. Concerning horizontal windmills and water- 
wheels, with oblique vanes.. 

Obfervations upon the~effeds of common wind- 
mills with oblique vanes, have led many to imagine, 
that could the vanes be brought to receive the direct 
impulfe, like a ihip failing before the wind, it would 

fce a rery great improvement i& point of power : 
while ofehers attending to the extraordinary and even 
unexpected effefts of oblique vanes, have Been led to 
imagine, that oblique vanes applied to wtfter-miUe, 
would as much exceed the common water wheels* 
as the vertical windmills are found to have exceeded 
all attempts towards an horizontal one. Both theffe 
notidns, but efpecially the firft, have fo plaufiHe an 
appearahce, that of late yeaft there has fcldom beeri 
Wanting thdfe, who have affiduoufly employed fhem- 
felvesto bring ttt bear deftgrts of thkkind: it may 
not therefore' be unacceptable to endeavour to fet this 
ttiatttr in a clear light. 

* Plate VI. £g 2d. Let AB be the &&*ow of a 
plain, upbn which let the vyittd blow in the dire&ieti 
C D, tvith fiich a velbeity as to deferibe a given fpace 
U % in a given time (ffcppbffe i feeond) ; and let 
A B be mewed parallel to itfetf, in the direc- 
tion C D. Ho W, if the plane A B moves' with tfefc 
lame velocity *s the Wind j that is, if the point B 
ihdvts thto' tfcefpate^E in the feme tit^e tfcat a 
particle of ait wt^tdd move thro' the feme fp&ee 5 k 
as plain that, hi this cafe, AeteCan be no preffureor 
iirhptrife of the wind upon the plane : but if tfee plane 
moves flower than the wind, in the- feme efire&ka*, 
fo that the point B may move to. F, while a particle 
*>f air, fetthig out ftttfn B at the feme mftaht, would 
move to E, then B F wSl eiprfcfe the velocity cif the 
plane; and the relative velocity of the wind and plane 
Will be exprfefibd by the iineVE. Let the ra«fe of 
• f E to BE be giVeh (ftfppofe 2 r 3.) i let the lint 
'A B rep rfefent thfe Trtipulfe of Ac vmA upon Ae plane 
A B, Wheh 'a&irtg'With its:i»htofle vetoeity B E 3 but* 

«* •* 





C 73 3 

* • * 

when adting with its relative velocity F E, let its im- 
pulfe be denoted by fome aliquot part of A B, as for 
inftance 4 A B : then will -J of the parallelogram A F 
reprefent the mechanical power of the plane $ that is* 
^ ABx^BE. 

idly, Let IN be the fedtion of a plane, inclined 
in fuch a manner, that the bafe IK of the redtangle 
triangle I K N may be equal to A B ; and the per- 
pendicular N K=B E ; let the plane IN be ftruck 
by! the wind, in the diredtion L M, perpendicular to 
I K : then, according to the known rules of oblique 
forces, the impulfe of the wind upon the plain I N, 
tending to move it according to the diredtion L M, or 
NK, will be denoted by the bafe IKj and that 
part of the impulfe, tending to move it according to 
the diredtion I K, will be expreffed by the perpendi- 
cular N K. Let the plane I N be moveable in the 
diredtion of I K only } that is, the point I in the di- 
redtion of I K, and the point N in the diredtion N (X 
parallel thereto. Now it is evident, that if the point 
I moves thro' the line I K, while a particle of air," 
fetting forwards at the fame time from the point N, 
moves thro' the line N K, they will both arrive at the 
point K at the fame time; and confequently, in this 
cafe alfo, there can be no preflure or impulfe of the 
particle of the air upon the plane I N. Now let I O 
be to I K as B F to B E ; and let the phtne I N move 
at fuch a rate, that the point I may arrive at O, and 
acquire the pofitioh I Q^in the fame time that a par- 
ticle of wind would move thro* the ipace N K : as 
O QJs parallel to IN; (by the properties erf fimilar 
triangles) it will cut N K in the point P, in fuch a 
#»anner, that N P=B F* and P fc^F E i -hence it 

K appears, 



appears, that the plane I N, by acquiring the pofi* 
tion O (^withdraws itfelf from the a&ion of the : 
wind, by the fame fpace N P, that the plane A B 
does by acquiring the pofition FG $ and consequently, 
from the equality of P K to F E, the relative im- 
pulfe of the wind P K, upon the plane. O Q, will be 
equal to the relative impulfe of the wind F E, upon 
the plane F G : and fince the impulfe of the wind 
upon AB, with the relative velocity FE, in the di- 
rection B E, is reprefented by 4 A B ; the relative 
impulfe of the wind upon the plane I N, in the di- 
rection NK, will in like manner be reprefented by 
•| I K ; and the impulfe of the wind upon the plane 
I N, with the relative velocity P K* in the direction * 
I K, will be reprefented by ^ N K : and confequently 
the mechanical power of the plane I N, in the direc- 
tion I K, will be 4 the parallellograro I Qj that is 
•j I JC x I N K : that is, from the equality *>f I KraAB*; 
and N K=B E, we foall have f I Q=| ABx|fi£. 
=4A B x £ B E=4 of the area of the paraUdlogom * 
A F. Hence we deduoe this 

General Proposition, 

That all planes, however fkuated y jhat intercept* 
the fame feSlion of the wind, and having the fame re- - 
Iqtive velocity y in regard to the windy when reduced < 
into the fame dire£tion> have equal powers to produce 
mechanical effe&s. 

For what is loft by the obliquity of the impulfe, 
is gained by the velocity of the motion. 

\ Hence it appears, that an oblique M is under no ■> 
difadvantage in refpedt of power, compared with a \ 
direct one j except what arifes from a diminution of 



it* breadth, ip refpftft to the fz&wn of (he wind : 
4hp breadth J N teiog by obliquity reduced *o I K. 

The difadyaottge of horizontal windmills there- 
fore docs not confift in this ; th«t each fail, when 
.dire&ty expofed to the wind, is capable of a lefs 
jpower, than zn oblique one of the fame dhnenfions ; 
jbuttha(t in m Jhorizqnfal windmill* little more than 
one fail c^n he a&ing at. once : whereas in the com^ 
moa windmill, all the fonjr ,a#: together : a&d therer 
fore, Aippofijig e»ch vane of a$ horizontal windmill, 
c£ the fiuaae -dimensions as each vane of the vertical, 
it is jj^anifeft the power of a vertical njill with four 
fails, wilj he /our tinaes greater than the power of 
the horizontal owe, let its number of vanes be what it 
will ; this 4ifa4vairtage arifes from jthe nature of the 
thiogj but if we confid$r tfce further difedwaage, 
thajtoarifes from tfte difficulty of getting the fails back 
again ^g^A the *?ind, &q. we *eed n<& wooder if 
this Jdnd of wiH is in reality found to have not above 
jorrj pf $he power of the common fort ; as ha* ap- 
peared in fame attempts of this kifld. 

In like manner, as little improvement is to be ex- 
pected from water -mills withidbKque vanes : for the 
power of the iamefe&ipQ of a stream of water, is 
not gceater when acting upon an ^oblique vane, ithan 
whe* aaing.^p«i a dWt one: and »ny advantage 
thatch he faade by intercepting 3 >gj?eater ft&ion, 
isfftiqh ^m^tiroesrit^fee doije in the, cafe -of w open 
xwer, wiU be ^aqnteibalJanced by tthe fupecior p efift> 
ajace, jhsit % h v*pfi6 dffpuid snee.t withiby 
*ight ^ngjqs $i> $hg cjirrent : whereas the qamoioo 
&¥*tft 4s^s'J»S¥e Wjiihifhe iWfttpriieady in&c fame 

K 2 Here 

[ 76 ] 

Eferc it may reafbnably be afkcd, that fihce our 
geometrical demonftration is general, and proves, that 
one angle of obliquity is as good as another ; why in 
our experiments it appears, that there is a certain 
angle which is to be preferred to v all the reft? It is to 
be obferved, that if the breadth of the fail I N is 
given i the greater the angle KIN, and the lefs will 
be the bafe I K : that is, the fe&ion of wind inter- 
fered, will be lefs : on the other hand, the more 
acute the angle KIN, the lefs will be the perpendi* 
cular KN: that is, the impulfe of the wind, in the 
direction I K being lefs, and the velocity of the fail 
greater ? the refiftance of the medium will be greater 
alfo. Hence therefore, as there is a diminution of the 
fedtion of the wind: intercepted on one hand, and ^n 
increase of refiftance on the other, there is fome angles 
where the difadvantage arifing from thefe- caufeis up* 
on the whole is the leaft of all ; but as thfe difadvan- 
tage arifing from refiftance is more of a phyfical than 
geometrical confideration* the true angle will beft be 
affigned by experiment 


In trying the experiments contained in Tab. Ill; 
and IV. the different fpecific gravity of the air, which 
is undoubtedly different at different times, will caufe 
a difference in the load, proportional to the difference 
of its fpecific gravity, tho* its velocity remains the 
fame ; and a variation of fpecific gravity may arife 
not only from a variation of the weight of the wholef 
column, 1 but alfo by the difference of heat of the air 
concerned in the experiment, and poffibly of other 
caufes s yet the irregularities that might arife from & 


C 77 ] 

difference of fpecific gravity were thought tb Be 
too (mall to be perceivable, till after the principal 
experiments were made, and* their effedts compared ;: 
from which, as well as fucceeding experiments, thofe 
variations were found to- be capable of producing a 
fenfible, tho' no very confiderable efied; : however, 
as all the experiments were tried in the fummer fea- 
ibn, in the day-time, and under cover ; we may fup- 
pofe that the principal fource of error would arife 
from the different weight of the column of the atmo- 
fphere at different times; but as this feldom varies 
above T t part of the whole, we may conclude, that 
tho' many of the irregularities contained in the experi- 
ments referred to in the foregoing eflay, might arife 
from this caufe ; yet as all the principal conclufions 
are drawn from the medium of a confiderable num- 
ber, many whereof were made at different times, it 
is prefumed that they will nearly agree with the 
truth, and be altogether fufficient for regulating the 
pra&ical conftru&ion of thofe kind of machines, for 
which ufe they were principally intended* 

V '' 




. y> 

•> / 





- I 

i »» 


.».- r : • •, 

* <> 


^ ^•^*- I 



. v ? • • • *